Oral evidence: Accessing public services through the Government's Verify digital system, HC 1748

Oral evidence, 18 March 2019.

Witnesses

John Manzoni, Kevin Cunnington and Nic Harrison.

Chair

Good afternoon and welcome to the Public Accounts Committee on Monday 18 March 2019. We are here today to look at Verify, the Government’s digital verification scheme. The National Audit Office has done an investigation into Verify because of some of the challenges in introducing it. It is a system whereby members of the public can verify their identity online through various third party providers to access services, including, for instance, HMRC and DWP services. Things have not gone smoothly, so that is why we are looking at it. Before we move into that, I want to welcome and introduce our witnesses. I alert Mr Manzoni that, as he might expect, I want to talk briefly about Interserve. Then Sir Geoffrey Clifton-Brown has a question on something before we move to the main session. From my left to right, our witnesses are Kevin Cunnington, director general at the Government Digital Service. Welcome to you, Mr Cunnington. I think it is your first time in front of this Committee. We are a friendly bunch. John Manzoni is the permanent secretary at the Cabinet Office, a regular attender here. Nic Harrison is the former director for service design and assurance at the Government Digital Service and is currently director of design, architecture and planning at the DWP. Mr Harrison, can you explain what being director of design, architecture and planning is? I don’t think you are an architect, so presumably this is digital architecture.

Nic Harrison

Architecture and digital strategic planning.

Chair

So you pick up the other side of the Verify system at DWP.

Nic Harrison

At the sort of macro architectural level, yes.

Chair

I am not sure if that’s justice or what. We will wait until we conclude. Mr Manzoni, obviously we have had some very big news about Interserve over the weekend. Do you want to explain where things are at from the Government end and how you are managing any outfall?

John Manzoni

The first thing to say is that Interserve is not Carillion and therefore we have been in dialogue with the company for some time. Their restructuring was a refinancing restructuring essentially. Had the shareholders voted in favour, it would have been a perfectly routine operation. In fact, what they call a pre-packed administration is a relatively routine operation in the event that the shareholders did not agree. The lenders, who were in control of the company all the time, had agreed between themselves what would happen on the other side of the pre-packed administration. Because the shareholders did not agree, the company went into administration on Friday afternoon and it came back out on Friday afternoon. So essentially what we have here is a transference of ownership of the company. Our concern, as you know, is always to make sure that public services are maintained. Public services have been maintained through this. We knew that they would be and we were also mindful and fully aware of the plan in the event that the shareholders did not approve. So at one level we supported the financial restructuring of the company because, as I said, this is not Carillion. We supported this one. Fundamentally, it is a good company. There were some issues associated with cash flow that got them in this bind. We still believe it is a good company. The company today has basically halved its debt and has now got £100 million worth of extra cash on its balance sheet, which it did not have on Friday morning. It is now in a better situation to serve Government and us as a customer than it was before.

Chair

To be categorical about this, there is not a single element of the service that they provide to the public that will be at risk as a result.

John Manzoni

I think that is a true statement.

Chair

It is a bold statement.

John Manzoni

I obviously can’t guarantee it into the future for ever and a day, but, as it stands today, there has been no blip.

Chair

One of the interesting issues, if we wind back to who the lenders are, is that Coltrane is a major shareholder, and it was their decision as a hedge fund to pull out that has caused this stage to happen. They are also a major shareholder in Capita and Mitie. Are you concerned about the fact that one major hedge fund has a very large stake in at least three that I am aware of—now two—major strategic suppliers to Government and about the impact if they acted similarly with the other two companies?

John Manzoni

This whole process was triggered because of the condition of the company, as opposed to any shareholder creating this outcome. They did create it in the last analysis, because they did not support the restructuring. As you know, Capita has been through a restructuring and is actually emerging in, frankly, a strong place. They are performing well, so we are less worried about that. I think we still need to continue to watch Mitie. It is not out of the woods. We are watching all these companies very carefully. This whole sector, as you know, has been problematic for some years.

Chair

You had a Crown rep on Interserve. Was it the same Crown rep all the way through? Has there been any interregnum or any gap between the Crown rep’s service?

John Manzoni

To be honest, I do not know the answer to that.

Chair

But you have a Crown rep now.

John Manzoni

We certainly have a Crown rep. I have been in dialogue with their Crown rep for six months at least.

Chair

Obviously we have been reading about this for some time. Although the moment was a surprise, we knew that there were issues.

John Manzoni

Yes, this has been going on for some time.

Chair

What about on Capita and Mitie? Do you have active Crown reps on those?

John Manzoni

Yes, we do. We are well aware of all those conversations, and where we believe there to be a need for us to get more heavily involved, we actually go inside the company. We have to be careful with Chinese walls. We put people on the inside of the company so that they can see the monthly cash flows of the company, so that we really understand the condition.

Chair

I think there are 27 strategic suppliers since we lost Carillion; there were 28. Is that right?

John Manzoni

They move around a bit, but it was of that order, yes.

Chair

Exactly—they go in and out, so there are roughly 27. Of those, would you like to tell us how many you are worried about at the moment that might either be going through a major change like this or are at risk?

John Manzoni

This sector, I would say, and it has always been thus—

Chair

When you say “this sector”, do you mean the service sector?

John Manzoni

I mean the service sector, especially where it is in conjunction with the construction sector. Carillion was thus. Interserve was thus. Mitie I don’t think has the same thing. That sector has been under some stress, and we have been watching those companies very carefully. You saw Kier do a restructuring and a refinancing, Capita doing a refinancing, and Serco a little before then. We have been watching those very carefully. We still consider this sector to be the most fragile, but there are one or two others that we need to keep an eye on. For instance—this is public—Ferrovial, which owns Amey, is another one. They have said that that sector for them is up for sale, so we are watching that quite closely. We are watching the companies. I do not have the list with me today, but we are watching the companies.

Chair

We have talked before a number of times about being too big to fail. Is not the danger that you have large companies, but some shareholders have significant shareholdings in a number of those companies, so there are various risks? If one of the service providers or an investor has a failure or withdraws, this could put a large section of public services at great risk. You can keep an eye on what is going on, but in a crisis does it not mean that the Government would have to step in—that, actually, they were too big to fail?

John Manzoni

It is difficult to answer that question in the general. In the specific, as you know, we took a very different line on Carillion from the one we took on Interserve. We were fully cognisant of what was about to happen with Interserve, and in fact supporting it, supporting the company and taking the steps across Government—because they have a lot of contracts across Government—to minimise the impact. It was about making sure that the Departments were not losing their nerve or worrying about what was happening in the headlines, because there were a lot of headlines, as you know, before it all happened. To the uninformed, that would be of some concern. We made sure that all that was going on. I am not going to describe exactly what we are doing, but we are taking different steps with different companies to make sure that public services are not at risk.

Chair

How are the living wills going?

John Manzoni

We have only just begun. That has been very well received, as you know.

Chair

But with Interserve, for example, if it had gone the other way and some collapse had happened, would you have known—

John Manzoni

We did have contingency plans.

Chair

So you knew every public sector service that they were running.

John Manzoni

Yes, we did.

Chair

And you had contingency plans for all of them.

John Manzoni

Yes, we did.

Caroline Flint

The GMB trade union is claiming that £660 million worth of public contracts were issued in the run-up to Interserve going into administration. When we last met as part of our strategic suppliers review I think we had an exchange about this in relation to Carillion. You made the point at the time that the Government has to be careful here, because you don’t want to mitigate, in a negative way or a too positive way, the fortunes of companies. You have mentioned, Mr Manzoni, discussions across Departments. Were there discussions with partners in Government about the risks of continuing to award contracts in relation to Interserve once it was clear that the company was in trouble?

John Manzoni

Just for the record, the £600 million that was reported in the press—today, I think—the bulk of that was actually awarded way before even the very first profit warning, in September 2017, for Interserve. The value of contracts with Interserve was actually running through from way, way before any hint that Interserve was getting into a tight corner. In the last 12 months, there have been awards to Interserve; I’m told it’s £175 million-worth of contracts in the last 12 months. Of course, as I have said before, we are not actually at liberty to pick and choose. Each time we make a contract, we have to make sure, or we choose to make sure that the company meets a set of financial tests. We can set those tests, or they’re routine and normal tests. And in fact, the last time the big contracts were awarded, the company did meet those tests. There was a late action that actually precipitated last Friday’s events, which happened after the last one of those had been awarded. What we are trying to manage here is this. How does a company look? Is it meeting its financial tests? If that looks good, we have no choice but to treat them the same as every other company, because we can’t pick and choose. Frankly, we knew that this company either would be successful with its shareholder vote or, if it was not successful with its shareholder vote, would do exactly what it did. We have known that for some time. There was a lot of noise from the shareholders. The shareholders have come out worse off as a result of Coltrane’s actions, as it happens, because they have ended up with nothing, as opposed to a little bit. We knew all that was going on. We also—I repeat—believe that Interserve is a strong company. It got itself into a financial crunch. From the public sector perspective, this is better today than it was on Friday morning, because it has more cash on its balance sheet and has halved the debt, and the lenders have stepped in, because they take the same view, and have said, “Actually, we’re going to back this company. We’re going to get it out of its financial hole and we’re going to take it forward.” So our judgments have been proven okay on this, I think.

Gareth Snell

Mr Manzoni, when we looked at Carillion last time, we asked some questions about the Crown rep system and how viable that system really was for providing Government oversight. You said a moment ago that you had known that this was coming for a while. Does the Crown rep system still provide the level of oversight and involvement that you actually need, given the precarious nature of, we now find, a number of these large infrastructure organisations and the value of the public contracts that they hold?

John Manzoni

In this instance, as an example, the Crown rep has been, I know, in a series of meetings with the company over a very long time, and at every meeting that I have had about the company, or indeed with the company, the Crown rep has been present, so they have actually been central to this particular example. And I am satisfied; he has done a sterling job on this. They have done—he has done; it is a he—a particularly good job in this case. They have been right on top of how this has played out and have, frankly, demonstrated good judgment in terms of “Should we do this? Should we do that?” and what decisions should be made, so I am satisfied. By the way, as you may know, we have recently launched, as a result of Carillion, the outsourcing review, which has a series of steps that will take place—11, actually—changes to how the public-private interface works. That is going to take some years, probably, to do, but will over time fundamentally, I think, improve the relationship between the private and the public sector. Between them, all those things are—I think we are in a much, much different place from where we were, say, three years ago, when it was sort of “price wins”. That is not the way to do this.

Sir Amyas Morse

I want to ask about the indebtedness. I remember that, with Carillion, the problem was that they had got involved in ventures in the Gulf and so forth that had not been successful, so they brought that into the public sector contracting situation. What was the source of all this debt on Interserve’s balance sheet? It wasn’t anything like that, I hope.

John Manzoni

No. The source of the issue for Interserve was a series of one-off energy-from-waste contracts that have been slow in coming to fruition. The company took on a relatively isolated set of issues as part of the construction, and it got into a crunch. The cash flow was not where it had predicted it to be, and we could see that that was a relatively isolated case. They spotted it quite a few months ago, and they were moving to isolate it so that the underlying business can now proceed. There are two or three chunks of this company, and that particular issue has now been isolated and sorted, we think. It was a particular energy-from-waste issue, which was in the construction sector.

Sir Amyas Morse

It does sound like they plunged in pretty deep.

John Manzoni

Into energy from waste—they made some big misjudgements.

Sir Geoffrey Clifton-Brown

My question involves two sides of the same coin, to try to get to the bottom of this. When Carillion went into administration, at that point the Government did not have a list of every single contract it held at central Government and local government level. Today you are telling us that you know precisely every contract that Interserve has. Does that include local government?

John Manzoni

That is certainly true for central Government—I was assured that contingency plans were in place. It probably doesn’t cover everything in local government, but it covered the majority of issues.

Sir Geoffrey Clifton-Brown

The point is that if it doesn’t include local government, how can your financial tests be accurate? If you agglomerate every public sector contract, the financial tests might come out with a different answer.

John Manzoni

That is a reasonable point. The financial tests are generally about the balance sheet strength of the company as a whole. Generally, there are a few big contracts. In Interserve’s case, there are 600, or whatever. There are four or five central Government Departments where contracts are really big, and those in local government departments tend to be smaller than that. There are multi-hundred-million pound contracts with the MOD for the provision of FM services from Interserve, and similarly with the DWP. The big ones sit in central Government. There will be a tail—we did not get all of them, but we were pretty comfortable that we had contingency planning activity and plans in the event that it did not go well, but frankly, we were confident that it was going to go the way it went.

Sir Geoffrey Clifton-Brown

May I suggest that in future for all these big services companies, you try to collate the entire contract position in both central and local government?

John Manzoni

Part of the next phase of the outsourcing review will get at that—living wills and such things. Now you cannot click your fingers and do that, because, as you know it, is a delegated activity. The information systems do not necessarily pertain, and neither do they in the companies, as we saw with Carillion, which had no idea what the contracts were worth.

Sir Geoffrey Clifton-Brown

May I ask a question about something totally different, which you may or may not be abreast of? As you know, I have a constituency interest in the defence fire and rescue contract. You will be aware that Serco put in a judicial review, which I have referred to previously in this Committee as a spoiler, because it failed to get the bid. There is an issue of public safety because Capita is not being allowed at the fire service college to continue with the contract, yet the MOD is not upgrading its training services. That is not a satisfactory situation. Can you bring us up to date regarding where it is at the moment?

John Manzoni

I can’t, except to share your view that it is not a satisfactory situation. I am not best pleased that a spoiler action has been taken by Serco. I think that is holding things up, and it is not good for the outcomes. One thing that we want to get beyond is the rather litigious approach to winning and losing contracts. Part of a more sophisticated relationship is that we get beyond that, which is important. I don’t mind saying in this forum that I think it is a shame we are in this position on this contract, and I agree that it is not helping us to move forward.

Sir Geoffrey Clifton-Brown

So while you might not want to share this with the Committee, do you have any plans, working with those big contractors, to reduce the litigious situation you are talking about?

John Manzoni

We are hoping to address it as part of the post-Brexit procurement regime going forward. Q25 Chair:When you say post Brexit—is that because it would no longer be under EU procurement laws?

John Manzoni

Correct—so that we are able to set our own expectations, rules and procurement environment.

Chair

So you have a draft or something somewhere, and people are aware of that in the Cabinet Office.

John Manzoni

Yes, there is. One of the objectives is to see if we can reduce the slightly litigious nature of people who lose.

Chair

Thank you very much. That is interesting. We would love to start talking about Eurotunnel, too, but we will be here all day if we do that. I will ask Gareth Snell to kick off on the main issue in question today, which is the NAO’s investigation into Verify, what has happened and where things are going.

Gareth Snell

Mr Cunnington and Mr Manzoni, could you give me any of the quantitative KPIs from any of the five business cases that Verify has actually met? Just one example of any quantitative KPI from any of the five business cases.

Kevin Cunnington

On the current business case, we are on track in terms of the volume. The most recent business case from 2018—

Chair

Can you speak up Mr Cunnington? Although this room is a modern room built for speaking, it is not actually that clear.

Kevin Cunnington

Sorry. On the 2018 plan, we are on track, in terms of the volumes of Level of Assurance 1 and Level of Assurance 2 verifications that we have.

Gareth Snell

Okay. Any others?

Kevin Cunnington

You are testing my memory now. I cannot think of any off hand.

Gareth Snell

Any from any of the previous business plans?

Kevin Cunnington

No. Generally the previous business plans all underperformed in terms of the volume of Verify customers that we expected.

John Manzoni

Avoided costs have not been far off, have they?

Kevin Cunnington

No. If we go back to the macro level—are we allowed to talk about the business case? In the 2016 business case we said we would spend £212 million, but in reality we have spent £154 million to date. Over the eight years of the programme, on a like-for-like benefit against those costs of £154 million, we had benefits of £366 million. Ultimately, the system itself covers its costs substantially.

John Manzoni

If I may help, the benefit-cost ratio, from the very first business case, to the state today is 2.4; it is exactly the same. That is met.

Gareth Snell

So that is three. I am just going to challenge you, if I can, on the cost. Obviously, the cost of £154 million was not due to you being any more officious in the application of the programme, was it? That was because the volume of usership was considerably below expectations. So you are saying that spending less, because you failed to meet targets, is now a success, in the Cabinet Office’s definition of success.

John Manzoni

Maybe I could reframe that. I do not think there is any question that Verify did not meet its original business case benefits. Nobody is sitting here saying, “Right, it met its projections.” But I think the issue is more one of hopelessly optimistic projections in the original business case than failure, because it is working today, as I have said to you. The benefit-cost ratio of Verify actually meets the original business case today. While the original projections were really way off—

Gareth Snell

I will come to the previous business cases, because you are right that they were significantly way off, and several reviews and a number of business cases still maintained optimistic trajectories—let’s say optimistic for the sake of diplomacy. To what do you ascribe the low take-up across Government for this system?

Kevin Cunnington

Of the 46 presented services, you will probably be aware that only eight of the original ones have been delivered by Departments, 11 new services have been brought forward, seven ceased to operate, 18 are not yet fully digitised and 14 do not require identity. Maybe I can just tell you the history of this, because I was in the Department for Work and Pensions as one of the people who submitted the original five DWP services that we thought would form part of the identity service. The first of those is universal credit, which continues to use Verify, of course. The remaining four never did, for various reasons—they weren’t transformed or they weren’t digitised—but two of the services, “get your state pension” and “check your state pension,” came forward. Candidly, that is pretty indicative of what was going on in Departments. In 2015, we had only 25 live services across the whole of Government, yet we were projecting to have 46 more incorporate Verify over the next few years. That has just not turned out to be true. The Government has not transformed as quickly as we had hoped; therefore we have seen this reduction in volume.

Gareth Snell

Without wishing to paraphrase you, Mr Cunnington—I will do it anyway—it is the Government’s fault that they have not transformed their service efficiently to catch up with your groundbreaking technology?

Kevin Cunnington

The strategic aims of Verify—not the KPIs; I probably should have started here—were fourfold. One was to create a standard in the marketplace. That is the most important thing. I think we can say that we have absolutely completed that.

Gareth Snell

It is quite an expensive standard though, is it not—£154 million for a standard?

Kevin Cunnington

Let me point out the other three benefits in our strategic aims. One is, on the standards front, that our standards have been accepted by Europe as passable within the European scheme and they have been acknowledged by the US as the foundations for their own standards. The second thing that we set out to achieve was a commoditised cost model, which we claim today that we have got to. We have got the price down to a point where it can be bought by the Government and by the private sector. The third strategic aim was to produce a safe and secure service across a multiplicity of private sector vendors, which again we have achieved. I contrast that with some other countries that have been breached; the UK never has. The fourth objective was the mass adoption of the service. Again, you are right to say that that is where we have not had the success we were hoping for.

Gareth Snell

Yes, so there was the application of the first three, which are qualitative, and then the one quantitative that was meant to derive actual benefit to the Government and the taxpayer, which has not actually been met. It is not a trick question—

John Manzoni

There is clearly an issue here where, first, the original projections were optimistic—

Gareth Snell

Why were they so over-optimistic?

John Manzoni

Because I think the Government at the time, back in 2015—or was it before then?—were optimistic about the level of transformation that could happen across the services. I differentiate between digitisation and transformation. There is no question but that that was optimistic, and it is not as if people were blind about it. There was a bit of us attempting to assert that we could build volume by telling Government Departments to use Verify. I fully hold my hand up that we were a bit slow in intervening in that, because I do not think that the incentives for Government Departments were helpful for it, for perfectly good reasons, because they had not transformed their services and because there was a start-up problem of cost. Their incentives were not to do that. We carried on for a bit too long saying, “Well, we could just tell them and they’ll use it.” That did not happen.

Chair

That has happened in other parts of Government as well. We have had this before with shared services.

John Manzoni

Yes, to some degree, it has happened. Therefore, we have to be a bit more sophisticated about how we enrol. That was absolutely a problem here. It is quite important to remember that the reason that we carried on as we did was that at every point, Verify was washing its face. It was not just costing; it was delivering more benefit than it was costing, so—

Gareth Snell

Well, let us for one second take that claim at face value. Yes, it was delivering more benefit, but—let us do the comparisons—the overall benefit that it was estimated to achieve in 2016, in avoided cost to Departments alone, was £466 million, but it came in at £148 million. On the idea that it was deriving benefit, it may have been deriving some benefit, but again, as with the projections for uptake, the amount of benefit that you were intending to derive was considerably smaller. In fact, it was £2.5 billion over 10 years and you came in at less than a quarter of that.

John Manzoni

What I actually said was that the benefits were in excess of the costs at every point—that is what I actually said, and that is true today. In fact, they are 2.4 times the cost.

Gareth Snell

The costs that are in here for the service talk about the direct costs incurred by Verify and the Cabinet Office. Throughout the Report, there are examples of departmental costs having been incurred to use the system that have not been consolidated to the costs that you say are being overplayed by the benefits. Has the Cabinet Office actually sought to quantify the entire global cost—not only to your Department, but to the Government as a whole—for the service of Verify, so that you could actually say that the benefits derived have been greater than the costs incurred?

Nic Harrison

I think the answer to that question is complicated by the fact that we provide Verify as an identification service for departmental services that want to use citizen ID. If a Department is transforming a service and needs to use citizen ID, it needs to do something. There are costs that transforming that service inherently carry, and the cost of using Verify is part of that. The benefits and costs for the actual service lie in those Departments. We do not claim—

Gareth Snell

That is just cost shunting, isn’t it?

Nic Harrison

Well, no; it is trying to make sure that the cost appears in only one place.

Gareth Snell

Okay. So there has been no attempt to quantify, even by looking at the consolidated accounts of each Department, what the potential costs and benefits of Verify have been to the Government as a whole?

Nic Harrison

In my tenure, we have not tried to separate Verify from other identity costs. By the same token that I do not know what Departments have spent on Verify, I also do not know what they have spent on other identity solutions. That is a decision for Departments.

Chair

Well, Mr Snell knows your exact unit costs.

Gareth Snell

Yes. We will come to that. Mr Manzoni, figure 6 in the Report details quite a nice timeline of various different decision points, including when contracts were signed and when business cases were completed; in one case it is business case version 6, which makes me question what happened to business cases 1 to 5. When you came to the board in 2015 and looked at this, what persuaded you that this was not just a money pit or an exercise that would not really deliver commercial benefits? What made you think at that point that the projections that you were told would be derived would actually happen?

John Manzoni

Verify went live in May 2016.

Gareth Snell

But the public trial started in 2014.

John Manzoni

The trials did, but the service—

Gareth Snell

The contracts were signed in 2013, and the business cases were agreed in 2014 and 2015.

John Manzoni

That is all true. It went live in May 2016.

Gareth Snell

When you saw the business case in 2015—

John Manzoni

I am not sure I was concentrating on the business case in 2015; I cannot remember, actually. However, I did get involved through 2016, when it became increasingly clear that we were not achieving the volumes that we were hoping to achieve.

Chair

So you are saying that you did not keep an eye on that business case, looking at the projected business—

John Manzoni

I cannot actually remember. I really started concentrating after the service had gone live in 2016.

Chair

When it started going wrong.

John Manzoni

It wasn’t meeting its projections. At that point—

Chair

Mr Manzoni, you have a reputation in this Committee of being a bit of a troubleshooter across Whitehall, trying to get the system to work together and to get the centre to encourage Departments to do things that are sensible, long term and cost-effective. You are saying that you can’t really remember that whether you looked at this business case?

John Manzoni

I admit that I might have been a bit late in the intervening. I intervened quite heavily in 2016. Indeed, through 2016 we were looking at it and we were working with IPA, and then we made a fundamental—

Chair

So who was looking at it? The volume bit, as Mr Snell is going into, is one thing that helped it stack up, yet figure 3 shows the number of Departments buying the service and figure 6 shows some of the other problems.

John Manzoni

I have said that I think this is a— There are multiple things here. No. 1, we were asserting that we could tell Departments to use Verify, and you are quite right that the incentives were for Departments not to use it, because it was too expensive. That was a start-up problem. No. 2, the services weren’t transforming as quickly as everybody had hoped, including all the Departments. The original business case in 2015 had an aligned perspective. The Department and GDS were saying, “This is how we are going to transform the activities,” so they didn’t transform as fast as they had. Was I six months too late in intervening? Quite possibly. I do believe, though, that when we did start intervening, we fundamentally changed it. It is quite complicated, because it is a service in the public domain with contractual partners and everything else, so it has taken some time. We can talk about where we are now, but once we really started to accept that the volumes weren’t going to be grown in the way that we thought—that took quite a long time, because it took a lot of discussions between me and the permanent secretaries of the big Departments—we finally used McKinsey as a review. Remember that, all the time, the identity market was adjusting and evolving in time. When we started this, it was brand new.

Gareth Snell

Just so I am clear—let me get the timescale correct in my head, Mr Manzoni—when in 2016 did you actually intervene?

John Manzoni

Well, Nic joined in October.

Kevin Cunnington

In September 2016.

Nic Harrison

I became an SRO in October 2016, having joined GDS just a month before. The first thing I did was look at Verify in some detail, and in November I wrote a detailed report that made 11 recommendations. I think my report said directly that Verify was not on a glide path to success, so we needed to intervene and do things. I made a number of recommendations, many of which we took into action.

Gareth Snell

Sorry—this is October 2016, so this would be the October 2016 business case re-projection? These are part of the 11 recommendations? Is that correct?

Nic Harrison

No, the 2016 business case had pretty much already been written at that point.

Gareth Snell

Okay, so what did you change from the 2016 business case for your 11 recommendations?

Nic Harrison

The 2016 business case had been written and had been through the Cabinet Office investment committee and the Treasury—it was a business case. As I came in, I recognised that it was unlikely to deliver in its current form.

Gareth Snell

Literally a month after it had been signed off and agreed by the Treasury, you said to the Cabinet Office, “This is not going to work.”?

Nic Harrison

I certainly said it was not on a glide path to success. I came in with my eyes open—

Chair

What were the Belisha beacons that stuck out and made you say, “This isn’t going to work.”?

Nic Harrison

It was massively behind on volume.

Chair

Was it that people were putting up the volume of users to make sure the numbers matched? The projected volume of users was increasing in the business to make sure the business case balanced? Was that what happened?

Kevin Cunnington

We just didn’t have enough adopting services. That was the bottom line.

John Manzoni

I don’t think the projection was changed. It was the continuation—

Chair

So it was over-optimistic from the beginning?

John Manzoni

Yes.

Gareth Snell

The 2016 business case projection was changed. Figure 2 shows that there was a deviation. There was a reduction in the projected uptake of users. As it happens, that failed as well. What I want to understand from Mr Harrison—this is quite important—is this. Figure 6 says, “Oct 2016…GOV.UK Verify Programme Business Case Version 10” agreed. You are saying that you came in just after that, looked at that business case and said, in your professional opinion, “Actually, this is not going to achieve the outcomes that the business case has identified, even though they have been signed off by the Treasury.” What did you do next? What happened next?

Nic Harrison

What happened next was that, after my report, which made 11 recommendations, we started to explore those recommendations. They were all measures to build more volume. It is as simple as that. It is a volume game. They were all ways we could potentially build more volume. The services that had been identified in 2015-16 were frankly the art of the possible, not the art of the art of the probable. They were services that could be transformed and could have a requirement for identity. We got into in-depth discussions with Departments trying to on-board these services. For a very good, sensible set of reasons, Departments had other priorities and other pressing issues. Services that were put on the original list in perfectly good faith by all sides just became impractical in a short-term timeframe. We started to look at other services and other areas where we could derive volume. We talked to local authorities and the health service, which was originally excluded from the business case.

Gareth Snell

What were your recommendations for the local authorities and the health service? As I understand it, no local authority has taken up this contract offer at all.

Nic Harrison

We continue to work with local authorities—

Gareth Snell

I am very clear in my wording, Mr Harrison: no local authority actually uses Verify at the moment, do they?

Nic Harrison

That is correct.

Chair

But in 2016 you were suggesting that that was an opportunity for growth.

Nic Harrison

Correct.

Gareth Snell

Can I ask about the timescales? In April 2016—I appreciate that that was before your time, Mr Harrison—contracts were signed with additional providers. In October 2016, the business case was signed off by the Treasury and the Cabinet Office. You came in after October and said that actually you had 11 recommendations. At what point did you have a conversation with the additional suppliers to say, “Actually, the business case that we now have, to service the demand that we have signed up for you to deliver, will not materialise”?

Nic Harrison

I never had that conversation in exactly that terminology.

Gareth Snell

Did anybody have that conversation, Mr Manzoni?

John Manzoni

Not as far as I know. I don’t know the answer to that. I think in some senses the context of this was complicated, wasn’t it—

Gareth Snell

Well, not really.

John Manzoni

—because they changed the management and leadership of GDS just before that. One of the reasons for doing that is exactly this sort of thing, so it is not—

Gareth Snell

When did you come in, Mr Manzoni? Just remind me.

John Manzoni

Into where?

Gareth Snell

Into the Cabinet Office, with responsibility for the project.

John Manzoni

August 2015.

Gareth Snell

Right. Bearing in mind that Verify’s original programme business case came in 2015, who agreed to sign off the contracts for additional suppliers in 2016 when there was no immediate evidence that there would be an increase in supply?

John Manzoni

I can’t remember.

Gareth Snell

You can’t remember. You are forgetting a lot of things today, Mr Manzoni.

John Manzoni

No. That was quite a long time ago. I came in—

Chair

Well, you did come to this Committee to discuss this.

John Manzoni

I accept that.

Chair

There is an NAO Report, which we note is uncleared—you did not sign it off with NAO.

John Manzoni

No, I didn’t, because it was issued before I got to sign it off.

Chair

Could the NAO just clarify that?

John Manzoni

Since you raised it, we were in a discussion, and I found out that the Report had been issued. There was not much point in signing it afterwards.

Chair

Ms Jones, would you like to comment on that?

Siân Jones

The Report was sent out from the CAG’s private office on the Monday—that was the final version. Previous to that, we had had two to three weeks of discussion with GDS, and then the formal Report went out on the Monday with a request for comments from Mr Manzoni’s office by the Friday, which did not appear.

John Manzoni

I cannot remember when we had our meeting, but we did not—

Sir Amyas Morse

Unfortunately, you cannot just go on and on letting it drift. We have made it pretty clear—

Chair

So there were a lot of discussions with your office, Mr Manzoni, and with Government about this Report.

John Manzoni

We had a meeting about it, and the Report was issued—that is fine.

Chair

We might go back to that meeting later, but let me bring in Mr Snell first.

Gareth Snell

Just a final point: you came in and made your 11 recommendations in October 2016, Mr Harrison, and then in October 2017 the Cabinet Office commissioned an external review of the strategic options. How many of your recommendations fed into that review?

Nic Harrison

I would think that all of them certainly featured in it.

Gareth Snell

So what significantly changed, in terms of the way in which you operated, between the business case in October and the outcome of the review a year later?

Nic Harrison

What do you mean?

Gareth Snell

Presumably you had a business case that you knew was not meeting expectations on volume. You made your 11 recommendations and an external body was then commissioned to strategically renew Verify. Throughout the entire iteration of this programme, there have been five business cases and 20 internal and external reviews. What I want to understand is at what point you actually started to change what you were doing in order to act on those reviews and try to improve the service.

Nic Harrison

The answer to that, as is often the case in these agile-type developments, is that we were making hundreds of changes throughout that period, all of which were quite small, because that is the nature of identity. The problem in identity management is that there is no one silver bullet; it is a combination of what data sources you have, what technologies you are using, how the physical user journey exists in the software, and how it is integrated into the services that want to use it. There is rarely one single thing that you can do that makes a really material difference. It is a question of going through the hard work. We were working very closely with the identity providers, looking at their data and looking at our data, and trying to improve how their services worked. They were themselves looking at where they could get new data sources and how they could improve their user journeys. We were working with Government Departments to try to expand the list of services that could use it. It was a continuous full-time job, without a single killer punch, I would say.

John Manzoni

If I may add to that, Nic’s review said we should work LOA1. As I understand it—Nic will put me right if I am wrong—that was launched by the DVLA in August 2017. LOA1 launched as a result of both Nic’s review and also the McKinsey review, which is the review you referred to.

Gareth Snell

You referred to it.

John Manzoni

Sorry, the external review that you referred to. That resulted in a substantial change to the strategic direction of this programme, which was formalised in the business case first in December and then in February 2018. Through the course of 2017, we were working with Departments to try to see if there was a different way of building volume in the context of their perfectly sensible decisions—as Nic has said—that they weren’t going to lean in and build the volume. We were working with Departments then. We ended up with an external review and that external review resulted in a substantial change at the end of 2017 and the beginning of 2018, in terms of the strategic direction of this programme.

Gareth Snell

Although the strategic direction of the programme may have changed, the uptake in the number of people signed up to Verify didn’t significantly increase at all over that period. Given that you had an external review and given that, Mr Harrison, you had had 11 recommendations, those recommendations clearly failed to meet expectations. Why did those recommendations fail? Even after the strategic shift in direction, what then prevented Verify from making the volume that was expected under all the business cases?

Nic Harrison

This is a volume game.

Gareth Snell

Yes and you are not getting any, so I am trying to understand why.

Chair

You have said that several times already, Mr Harrison. Volume was the key—we have got that—but you didn’t get the volume right.

Kevin Cunnington

Do you mind if I help out? Our expectation was that Departments would transform more quickly and services would come back more quickly to where, if you like, I signed off those original five programmes in DWP, but only three happened. It just wasn’t happening. We recognised that it wasn’t happening. As John says, the key point for us was when the McKinsey review highlighted that Government were only going to ever be 2% of citizens’ use of digital identity and the vast majority of their use would be in the private sector. That is when we began to actively plan: the Government are going too slowly, we are not going get there, so how do we need to make this work and how do we get commoditised pricing at scale?

Gareth Snell

I want to come on to the commodity pricing of it and the commercialisation of data, but I want to talk about the user experience for a second. The original business case stated that they wanted the verification success rate to be 90%. Obviously, in those early days that fell well short. How much do you think that impacted on other Departments wanting to take on that service?

Nic Harrison

There are a couple of things in there. The 90% figure appears in the 2015 business case, which is highly speculative and shows best probabilities. It was derived from what the likely success rate could be, based on demographics and looking backwards from a period when the product had been a success.

Gareth Snell

It is also the figure that was in the Government business case that was used to allow sign-off of this programme to go forward. Regardless of whether that is a figure that is right or not, it is the figure on which this business case was predicated, and it failed to meet it.

Kevin Cunnington

No question. It was adjusted down to the range of 20 to 30 in the 2016 business case, when we recognised that wasn’t going to be true. As Nic says, the 90% was based on a future state where a lot of people are already verified—not the challenge of signing up. Nic uses a very useful analogy around the challenge in signing up for a bank account versus logging back into your bank account. It is important to recognise that 90% was based on a point at which most people had already got their bank account—their verification ID—and were signing back in, not trying to create it.

Gareth Snell

Mr Cunnington, those services that were not using Verify—what were they using instead?

Kevin Cunnington

Often nothing—or most often nothing. So, of the original 46 that were planned, only eight of them actually went ahead to use Verify. Seven ceased to operate—they just disappeared; they weren’t good ideas—18 weren’t transformed at all and 14 didn’t require identity. A good example of the latter category would be a benefit with a medical assessment. What we thought way back in 2015 was we would identify people as part of the process; but what we recognised is, since we had got to do a medical assessment anyway, you know full well the person in front of you is the person who is in front of you. So there is no need to make the system more complicated.

Chair

You only discovered that later.

Kevin Cunnington

Yes, that’s true. We only did discover that later.

Chair

So what was the business case—were you modelling any of these pathways for the user?

Kevin Cunnington

No, not explicitly. Not in 2015, when the original business case was put together. It was a list of what are the most likely digital programmes to use Verify—

Chair

It seems a bit “finger in the wind”, when you have got all these things that might need Verify, but are not actually saying to the people using those processes, “Would you need Verify?”

Kevin Cunnington

Yes, it’s true, but of course you have to have perspective on this. The Government only had 25 live digital services at that point, so it wasn’t as if the Government had large amounts of expertise in assessing—

Chair

Evidently not.

Kevin Cunnington

No, it’s absolutely the case that it wasn’t.

Nic Harrison

In fairness, this list of services was put forward by the Departments. This was not something entirely created by GDS. This was a list put forward by Departments: “Here is a list of services that we believe, in a future online state, might need an identity element”.

Gareth Snell

Can we just look at figure 3, very briefly, while we talk about this? The systems that weren’t using Verify: how many of them were using the HMRC Government Gateway system?

Nic Harrison

I don’t know the breakdown.

Kevin Cunnington

I think it was six.

Chair

Six—which included HMRC.

Kevin Cunnington

It was almost exclusively HMRC—and, actually, DWP Check Your State Pension.

Gareth Snell

So given that we had a pre-existing system owned by HMRC, through Government Gateway, what prompted the Cabinet Office to think that reinventing the wheel, to have a new system, was the way forward?

Nic Harrison

Fundamentally, the Government Gateway system is not a citizen identity system. It is a secure log-in into a service. To be honest, the difference between those two things has bedevilled this discussion for as long as I have been involved in it. A citizen ID system is a way for a person to project their identity legally to a service without needing to be there. So that means someone who is using an identity system, it meets a standard—and we have set out a number of different standards within Verify: the LOA levels, or level of assurance. Level 2 has a certain meaning. It means it is enforceable up to civil court standard; whereas a secure log-in, which is what the Government Gateway system is—all that proves is that a log-in exists and you know the password to it.

Chair

Actually I beg to differ, having had the delight of using both. If I were careless with my paperwork—my numbers, or where my passport was—someone could probably log in with my Verify as easily as they could with Government Gateway. That might be easier, because Government Gateway is a very long number you have to remember, which I could never find.

Nic Harrison

That might be true up to a point but, over time, for example, one of the things we are buying from IDPs—the point of having this service—is that they are continuously checking these attributes are still true, whereas Government Gateway is always a point in time, of what you set up.

Chair

Do you mean the new pass code arrives to your personal phone, for example?

Nic Harrison

No, that sort of two factor authentication exists in both systems, but it is the level of background checking that goes on in Verify that makes sure that if your circumstances change it updates that regardless of whether you tell anybody or not. So if you move, for example, Verify knows about that and, therefore, will force potential higher authentication levels when you use it again.

John Manzoni

This is quite a fundamental issue, which, as Nic said, has bedevilled this project from the very beginning. These two things are very, very different. A Government gateway is very different from a digital ID, which is ubiquitous and can be used in many different places. The latter is much harder to establish, as we have found out, but it is a very different issue. It gets a bit technical and a bit complicated. By the way, since 2014 and 2015 the market has developed so that there is now something of a clamour for a digital identity, which there wasn’t when this started.

Siân Jones

Our understanding of the Government Gateway is that it isn’t just the initial log-in. I appreciate the point about apples and pears that we are talking about here. Clearly, beyond the initial log-in, HMRC is also doing other checks with the other existing information that it holds for its customers, to do checks.

Chair

That is HMRC-specific, isn’t it?

Gareth Snell

Looking at figure 3 of the NAO Report, on page 14, and putting aside the eight services that use Verify only, of the 11 that are there that have an alternative online access route, the HMRC services in particular—taking your point, Mr Manzoni, at face value, that the Government Gateway log-in is not the same as Verify—does HMRC incur a cost for registering a Government Gateway in the same way that the Cabinet Office incurs a cost for registering with a Verify account? Is it necessary to have a Verify account to undertake all the work that you have to do in those HMRC portals?

Nic Harrison

There is no requirement to have a Verify account if you use the alternative access method.

Gareth Snell

Does HMRC incur a cost?

Nic Harrison

They do.

Gareth Snell

Do you know what that cost is?

Nic Harrison

We have never been able to determine exactly what it is because, obviously, it is slightly an operational cost and is buried in other development costs. But there is an operational cost.

Chair

They are not paying a fee, are they?

Nic Harrison

No, they are not paying a fee.

John Manzoni

That is part of the problem with the earlier question, of identifying the true benefits and the true costs, because the costs of the Gateway are so embedded in the normal operations of HMRC, whereas the costs of Verify are necessarily explicit because it is contractualised. That is the fundamental complexity of trying to—.

Gareth Snell

But in theory, if you are somebody—I would say me, but I don’t use most of these systems—who has both a Government Gateway and a Verify account, the Government has incurred a cost twice for you to access the same service potentially through two separate routes.

Nic Harrison

Correct.

Gareth Snell

That is not value for money, we would all agree, I hope.

Nic Harrison

Agreed.

Gareth Snell

Okay. Can I briefly go back to the experience? Paragraph 1.17 on page 16 of the Report talks about the success rate for the numbers of people who attempt to sign up. Putting aside the 2015 figure of 90% and recognising that your current figure is 48%, Mr Cunnington, can you explain why the success rate does not actually cover how many people get through to the service they want in the end?

Kevin Cunnington

Yes, I can. GDS measures only the completion rate for people entering the Verify service and leaving it, because that is the only part of the process we have control over. The reason we do not extend beyond that is the privacy regime we are signed up to with the privacy groups. So we can only see end to end across our service. We cannot see which services call us, and vice versa, services cannot see which IDP we use. It is all part of the privacy model—

Gareth Snell

Okay. Once somebody is in the system, basically it is up to them whether they get to the service they want, but as long as they get into the system that is what you tick off as being a successful registration.

Kevin Cunnington

Yes, typically a citizen would type into Google, “Check my state pension”, let’s say. They would then go on to a set of Government websites that would enable them to choose their identity provider. They go through the identification process, which I measure, and are then matched back into the consuming service.

Gareth Snell

Okay. Given the limited access that you could get through Verify at the time, and partly linked to our previous conversation about Government Gateway, which I now accept are two separate systems, on the underperformance of Verify, coupled with the complex nature of its existence, how much of that was a factor that did not allow HMRC to take up Verify and instead chose to stay with Government Gateway?

Kevin Cunnington

There are two answers to this. There is a very simple answer, which is a learning, and then a very much more complicated answer as to why it is the way it is. The simple answer that we have learnt over time is that it is much easier for new services to work with Verify than it was for pre-existing services like the tax platform or universal credit, and that is just a learning. They were designed in a certain way. They are not so compatible with the way Verify works, and they would have to go back and do some rework, which they do not want to do, to make it work for the Verify system. As for the complicated reason why the Verify completion rates vary, there are three elements. One is that it depends on the level of assurance you want to get to. We know that the higher level of assurance on LOA2 is harder to achieve because, in your case, you need three specific items of evidence: your passport, your driving licence and some financial information. That is harder than getting into LOA1, which is relatively easy. In our best performing services today, our best performing LOA1 service at 95% is the health service pension. Our best performing LOA2 services are in the range of 70% to 80%. So first it depends on the level of assurance you want to get to. Secondly, it depends on the demographic of the user. Universal credit is particularly at an end of a continuum where we know people on universal credit do not have strong digital footprints. They might be using mobile phones to verify themselves rather than tablets or PCs, and it varies according to demographic. The third element is how strongly you nudge people into using Verify. A good example of this: for universal credit, absolutely do not try and nudge people into using Verify. The primary policy objective is to get claimants paid, and claimants go all the way through the process before saying, “Wouldn’t it be nice to verify yourself?” So you have always got three factors and that is why the rates vary: level of assurance, demographic of the user base and then the nudge behaviour that is in the consuming system to get people to verify themselves.

Chair

None of that is rocket science.

Kevin Cunnington

Are you saying it is rocket science?

Chair

It isn’t. There was a whole ID cards programme before this that could have told you some of that, yet you still didn’t put it in in the right volumes.

Kevin Cunnington

I am not sure I have understood your question.

Chair

It was just that you were way out on the volumes, despite knowing, as you have now outlined, that there were these barriers. You could have done an analysis of the demographics and an analysis of the digitally connected and those who weren’t who were using these services. Did you do any of that?

Kevin Cunnington

Yes, we did.

Chair

And yet you still came up with these high volumes.

Nic Harrison

All those original projects predate any of our time in this programme.

Chair

But you then walked into it. When the millennium dome was created, every time there was a problem they would just think, “In that case just increase the visitor number projections.” If you see it is wrong at that point—you obviously brought up your 11 key points, but did any of you at any point think, ”Maybe we need to do more than a reset, like a big, big reset on this or we have got to do some heavy lifting”? Mr Manzoni, you talked about trying to persuade Departments to buy into this and having lots of heavy discussions with Departments. I know this is a frustration that filters through your job, but where were you getting with those conversations?

John Manzoni

They took place through 2017. That is when they started taking place. As I mentioned, context is important. We changed the leadership at GDS. The leadership of this programme was changed. The new leadership took a view at the end of 2016, and then through 2017 we were increasingly aware that the volumes simply were hopeless. So we were in discussion trying to figure out what really was the incentive structure for the Departments. Why weren’t they using it? We were trying to get at the full and true costs of the Gateway process. We were trying to understand them and compare them with the Verify process. We were trying to do all that through the course of ’17. Ultimately in the summer of ’17 we decided to invoke an independent review, which then basically created a set of conditions. Quite frankly, it was a sort of negotiated way of reaching an objective. The truth is that this had driven everyone into their corners. That is what happens over time, and—

Chair

That is a polite way of putting it.

John Manzoni

Well. Therefore, it took time to get everyone back on to the same page, and the independent review helped that—

Chair

Cutting through that, as you are a direct speaker, you are saying really that the previous GDS—we will throw them under the bus as they are not here—had not really tuned into what users, in this case the Department users, wanted, needed or didn’t have the confidence—

John Manzoni

You saw some of the same behaviour with the shared services—you raised it yourself. There was an instruction that said you would use it—

Chair

And rural payments.

John Manzoni

Right, and it didn’t work. As we have just hinted, this is a really complex thing: it is quite technical; it is not the same as Gateway; and it has third party contracts. So this was really quite a complex environment in which to make a switch of direction and to acknowledge that the incentive structures just weren’t set up. You couldn’t carry on beating people up and saying, “Would you please use the service?”, because it wasn’t right. That is why it took the best part of nine months or something to get us heaved around into a new direction.

Gareth Snell

Presumably then, going back to figure 6, in February 2018, “Digital Identity Full Business Case Version 1 Draft 5”, that is the first time that the recommendations from Mr Harrison, the outcome of the 2017 external review and the three options all came together to give an idea of where this project should go next—

John Manzoni

And the OBC in the December of the year before.

Gareth Snell

If I had been an onlooker, an outsider peering through the windows of the Cabinet Office in that February—

Chair

That is what we are really.

Gareth Snell

I don’t spend too much of my time leering through windows. What would I have seen change in how your office approached this contract, in terms of the tangible conversations you were having with Departments, of the actions you were taking and, going back to my earlier point, of the conversations that you were having with contractors to make them aware that the volume was about to be rescaled down much below what the original contract values were?

John Manzoni

Let me start at the high level, but it was pretty fundamental. First, we had opened up LOA1 by that time, and previously we had been holding LOA2 as a high pinnacle that had to be achieved—so opening up LOA1 was the first thing. Secondly, we capped, and were explicit in the cap about the amount of money that the Government would spend from that point, and we invited the contractors into a wholly different commercial arrangement with Government in bidding for volume—so we were explicit in saying, “We are not going to continue to subsidise this. We have a level of money that we will put in. Do you want to play in those circumstances?”

Chair

With all of this, you are probably getting exasperation from me and Mr Snell. You would have just closed down other projects like this. Why didn’t you with this one? Is it because you were so tied in? What was the problem stopping you? GDS would have done the same—rural payments is a case in point—and tried to change it from the other way.

John Manzoni

It is not obvious to me that—With the programme as currently structured, if you buy its original objectives, which were to create an identity market for this country, which by the way increasingly outside commentators are saying now—

Chair

Funny, isn’t it, how identity cards come back one way or another?

John Manzoni

Commentators are increasingly saying we need an identify market. What have we done? We have created a set of standards, which by the way is internationally acclaimed and agreed—which is now with the outside world—and we have a set of providers, a market and a commercial contract—

Chair

Arguably, the banks—this is what—

John Manzoni

We will come to the banks, if you like. Right now, we have a set of providers who have bid in and said, “Yes, we’re prepared to play that.”

Gareth Snell

But, Mr Manzoni, you spent £154 million to get a series of standards, so I would hope that they are internationally acclaimed given the money that you spent on them, and you have providers, but under the renegotiation of the contracts in 2018, providers dropped out, so you have a shrinking pool of providers for the Verify service who only guarantee £21 million between now and 2020. At the end of 2020, which we will get on to in a minute, we actually do not know what will happen to this service, the cost price or the issues that have been discussed so far. I appreciate that you are trying to put a gloss on this, but you understand that, from a value-for-money perspective, £154 million plus another £21 million to have at the end of it no guarantees whatever is how you can read the Report.

John Manzoni

You can, if you choose to read it that way, but you have forgotten the £366 million of benefit that went with the £154 million cost.

Gareth Snell

Again, Mr Manzoni, the NAO has quite rightly said that it cannot work out how you have calculated that. It cannot replicate that level of benefit saving from the information that you have given it. Perhaps you would like to enlighten the Committee by explaining how you have quantified that level of benefit to the public purse, given that the NAO—the national auditors—appears unable to do so.

Nic Harrison

The answer is that the numbers that appear in various businesses cases that outline the costs and benefits are all constructed in the normal methodology that has been through the Cabinet Office investment committee and the Treasury. They are largely things like costs avoided, which is difficult to measure. I am not surprised that the National Audit Office cannot measure them independently, particularly the failure to deliver them as opposed to their projected size. We did not invent a new methodology for the benefit analysis.

Chair

It seems to us that the volume was always going to be a problem. You are all saying that it was before your time, but you are here to answer for the Government now. Why were they so far off? As I said, Mr Manzoni, is it this barrier where Departments do what they want regardless?

John Manzoni

No, I think there were two things going on. To go back to the volume point: back then, there was all sorts of optimism around what could be achieved and how quickly.

Chair

There is always optimism—that’s the problem.

John Manzoni

Well, I accept that, but the world moves pretty fast. We know a lot more. When I first started this job, I was in a conversation about universal credit and how optimistic that was when it first started. We had to reset that and today—five years later—it is a very successful digital service, but it took a lot longer than we thought it would at the time, so there is a whole contextual piece that the whole system was rather optimistic about the pace of digital transformation. That is not the case. Increasing realism, of course, makes it difficult to sit here and argue and justify, but back then, there were lots of examples of where we thought that we would go faster than we actually have done. You can digitise services, but it is a lot harder to fundamentally transform them to enable the Verify access.

Gareth Snell

I want to touch on the contracts that were brought in in 2018 and take us through to 2020. For the existing providers, you said, “This is the amount of money”—in your own words—“Do you want to play?” What benchmarking did you use to determine what was an appropriate level of remuneration in terms of individual transactions?

Nic Harrison

Specific benchmarking? There are some specific contracts that already exist that give us some comparators. Frankly, the principal story is that we have been in conversation with the IDPs themselves and understood some of their cost structures. We had some independent benchmarking, as part of some of the independent reviews, which gave us a good idea of what the unit cost could be. You have to bear it in mind that the costs across the individual suppliers are very different, so we were making a best estimation of what we believed would still be profitable. Obviously, when we let those contracts, we wanted IDPs to commit and play; there had to be an element of profitability so that they would still join the party, but we wanted nevertheless to limit the amount of money that we were putting up in a subsidy. It was a finely balanced judgment between keeping them in the game and trying to achieve the lowest price possible.

Gareth Snell

Paragraph 2.3 on page 18 says that the average overall price paid for new LOA2 verification was £22. Are you able to tell me what the price is in the current contract?

Kevin Cunnington

We are able to write to the Committee and tell you what that price is, but because it is a commercial arrangement, the commercial providers obviously do not want it in the public space.

Chair

You agree to write to us privately on that?

Kevin Cunnington

Yes.

Chair

Okay. We will wait for that, and maybe come back to you.

Gareth Snell

Is it substantially lower than £22?

John Manzoni

The other key difference was that we have said, “You can go away and build volume. You can go and verify citizens yourself.” That was another key shift of this strategy, and that means that they are now prepared to invest, and there are signs that they are beginning to invest, in building their own volume, which of course will—

Gareth Snell

Was it significantly lower than £22? Yes? Okay—does that mean that the original contracts were overvalued, and not good value for money?

Kevin Cunnington

It was a different point in time, I would say. Part of the early phases of the identity programmes was to try to encourage people even to be in this space. If you go back to 2015, it probably was not as obvious that it was going to be quite the market that it is today. To John’s point, we see very clear signs of real activity in the market by the IDPs today.

Gareth Snell

Okay. Where are they going to succeed in building volume where the Cabinet Office failed?

Kevin Cunnington

In the private sector, bearing it in mind that the private sector is 98% of the citizens’ usage of their identity. Government was always going to have a slightly cameo—

Chair

So you are very confident that they are going to keep investing?

Kevin Cunnington

They are very confident that they are going to keep investing. I have met with all the IDPs recently. We hosted a roundtable with all of them as recently as a couple of weeks ago. They all have very ambitious plans—

Chair

And the brand is still going to be Verify?

Kevin Cunnington

Yes.

Gareth Snell

You are confident that they will build volume, and that that increase in volume will ultimately bring down the unit price per verification.

Kevin Cunnington

I am confident that they will build volume. The price point is guaranteed in the current tiered contracts that we have as Cabinet Office with the IDPs.

John Manzoni

That is the strategy. We are not going to promise the strategy, but that is the strategy, which is a fundamentally different strategy from telling the Departments that they are the ones who are going to build us volume. That is why it is fundamentally different.

Gareth Snell

Two points on that, if I may. When this is spun out as a good commercial operation, if you find that commercial providers are suddenly making significant profit on the back of the commercial operators in the private sector, have the Government done any work to protect their intellectual property, having put the money up to start the platform in the first place? Is there any dividend repaid to the Government for this £154 million project that you have given to the private sector?

John Manzoni

It plays into the £366 million of benefit, but that would be a nice problem to have.

Gareth Snell

That is a no then.

Chair

That is not an answer to Mr Snell’s question.

Gareth Snell

What actions have you taken in the Cabinet Office to protect the intellectual property and the investment of the taxpayer into the Verify service once it is spun off and used commercially in the private sector?

John Manzoni

We will only pay what we are prepared to pay for Government services.

Gareth Snell

That was not my question.

Chair

Try again. Third time lucky, Mr Snell.

Gareth Snell

There is an echo in this room. Mr Manzoni, what action has the Cabinet Office taken to protect the intellectual property of the platform created through public money once it is spun off and used commercially in the private sector?

John Manzoni

I am not aware, but I think that is an ongoing discussion.

Nic Harrison

I think there is a slightly more complicated answer to that question, which I will try to give. There are two separate parts. The verification being done by the private sector on our behalf, which we are paying them for, is their IPR anyway. All those systems belong to them; they have developed them, and they have invested. That is what they have already paid for. There is an element of the Verify system—the hub and certain data platforms—that belongs to Government currently.

Chair

And the brand—the name “Verify”.

Nic Harrison

Yes, and the name Verify. Between now and the end of the existing contracts, it is our intention to, as you say, spin that off into the private sector. The commercial arrangements under which that will be done have not been finalised. It is possible that we can generate revenue. All things are still possible.

Gareth Snell

I am going to take from that—correct me if I am wrong—that there is currently no active pursuit by the Cabinet Office of any revenue income stream from the platform that £154 million worth of Government money went into, for a name brand that is owned and created by Government, and for a service that we are then going to have to pay to access commercially.

Nic Harrison

By the same token, we are not giving it away, either. This has not been decided. The method of disposal is currently under discussion.

Gareth Snell

So what is going to happen in 2020 when the existing contracts expire?

Kevin Cunnington

We and the Crown Commercial Service are currently in negotiation with the IDPs about the successor programme—the broad 2020 programme. If I just go back a bit, though, the strategic objectives of Verify were to set the standards, provide commodity pricing, provide a safe and secure platform, and get mass adoption. At some level, we will have achieved those.

Chair

But not mass adoption.

Kevin Cunnington

Through the private sector, we believe we will.

Gareth Snell

Sorry, but that is a very big statement to make, Mr Cunnington—that you believe you will. What is it that gives you that confidence when the last seven years of programme work has not delivered the volume that you needed, and as Mr Harrison has said, it is a volume game? What makes you more confident now that the private sector is going to achieve that where Government have failed?

Kevin Cunnington

There is more volume in the private sector. Some 90% of our usage with citizens is with the private sector, not with Government.