Examination of Witnesses (Questions 20-39)|
WEDNESDAY 22 MAY 2002
20. But part of the pattern would be assessing
the risk at that stage.
(Mr Gaskell) Yes, it would.
21. Do they assess risk at different levels?
(Mr Gaskell) Yes, they do.
22. How many gateway reviews did this project
(Mr Gaskell) This project has had three gateway reviews.
23. What levels of risk came out?
(Mr Gaskell) If I can just deal with the last one
which was done earlier this year, they recognised and confirmed
our own assessment that this was a high-risk programme, that it
was still achievable but that it would require a considerable
degree of effort on both sides in order to try and make that happen.
They also recommended that we should delay the decision until
the latest possible moment to give EDS a fighting chance to try
and bring the system on line.
24. You said there were three gateway reviews;
did the two earlier reviews have high-risk ratings as well?
(Mr Gaskell) Yes. From the outset of this programme
that has been the case.
25. So you have been sitting on what, by the
gateway process, has been assessed as a high-risk project all
the way along?
(Mr Gaskell) Yes. Can I just add to that, because
one of the things they do is look at the approaches that were
taken to managing those risks and, in particular, the contingency
measures we put in. On each occasion they have confirmed that
the steps we have identified to manage those risks have been the
right and appropriate ones.
26. So when did you consult the external IT
consultants? At the beginning or right at the end?
(Mr Gaskell) We have had a degree of IT expertise
within the Department throughout the programme. The particular
expertise I referred to earlier we specifically asked for to look
at the pattern of defects. So this was new expertise we were bringing
in. We have had a new chief information officer appointed and
one of his first jobs was to bring in some of that expertise and
to help the Department with the modernisation programme as a whole.
It is part of that resource that has been ear-marked to help us,
with EDS, to understand the pattern of defects and the pattern
of testing and what conclusions can be drawn from it.
27. Were those external consultants involved
in all or any of those three gateway reviews?
(Mr Gaskell) No, no, these are people that have been
brought in in the last two to three months to help us make that
assessment of the IT. I should add, for completeness, I have mentioned
three gateway reviews. However, because there was not a clearly
established pattern of gateway reviews, based on previous recommendations
from the Public Accounts Committee we have also had a series of
external and independent reviews to the programme before the gateway
reviews were set up properly. We stopped those when the gateway
process kicked in properly. That is when we had the two further
gateway reviews that I referred toone late last year and
one early this year.
28. What happens now?
(Mr Gaskell) What happens now is that we need a clear
and agreed re-plan from EDS to confirm how much time they believe
they need to complete testing. We then need to assess, jointly
with them (which we are doing) how achievable that plan is, how
much time that will involve, how much time we will need and the
approach we will need to complete the user-acceptance testing.
29. Do you know with any kind of certainty how
long it will take for completion of testing and roll-out commencing?
(Mr Gaskell) I think that would be a judgment that
Ministers would have to make, based on advice. There are a number
of options. One is that you could have a significant period of
elapsed time between the IT being ready and cleared user-acceptance
testing, or there could be a protracted roll-out period before
we go on to the new scheme or you can go for a shorter period.
It depends on what the quality of the IT is at that particular
30. But it would be months not weeks?
(Mr Gaskell) The gap between IT go-live and A-day
itself, if the quality of the IT is good, can be very short. Indeed,
there have been many projects in both the public and private sector
where there has been coincidence of the go-live and main implementation.
31. Within weeks?
(Mr Gaskell) There have been some where it has been
absolutely coincident. I would not recommend it.
(Mr Smith) It is worth dwelling on the tasks that
we need to undertake once we have the IT system. We would then
wish to do a proper user-assurance test to make sure that the
product that EDS say they are delivering, deliversand we
have got a lot of people working closely with thembut we
need, from our perspective, to make sure that it is up-to-spec.
So we need some user-acceptance testing. We also want to understand
how the IT system properly supports our business processes. It
is all very well writing this down on a piece of paper and drawing
models and pictures and having understanding of it, it is actually
when you start using it and saying "Here is a maintenance
application, here is how we work through the process using the
maintenance application"in a real environment rather
than a testing environment - that you start learning the real
lessons of how it will work in practice. Then we need to ensure
that our training packages for staff reflect the way that the
IT system will work in practice to support our business process
rather than how it will work in theory. So there is a degree of
effort and work that needs to go on.
32. There is a step between the completion of
testing and the roll-out, which you have just described. How long
will that process take?
(Mr Smith) As has been stated, it is entirely a product
of what the quality of the IT system is that is delivered. High
quality can mean a relatively short period. If there are still
issues within it then it may be a longer period. We are not going
to be in a position to have a judgment on that until we are far
closer to having the IT system in our hands.
33. I just wanted to go back to something I
think Mr Gaskell said, that EDS had not given you an indication
of when they anticipate finishing testing.
(Mr Gaskell) Yes, they need to give us that.
34. They have not yet?
(Mr Gaskell) They have given us an indication when
that is and we need to work around that, first of all, to confirm
with them that that is the assumption of risk associated with
that to see whether those are ones that are acceptable and achievable.
35. So in the two months since the Minister's
statement came out on 20 March you have not got it sorted as to
whether EDS's anticipated date is something you think is credible?
(Mr Smith) We do not currently have a date that we
believe we can say this is when we expect testing to be finished
because there are a significant number of issues still in testing
that need to be resolved now. However, until we see whether EDS
are actually proceeding in accordance with the plan and whether
things are moving on that elapsed time, it is going to be impossible
to say "Yes, by this date we are going to draw a line and
say the system test is finished" and we move on.
36. It seems EDS have not yet given you a date
so that you can then assess whether you think that date is realistic.
(Mr Smith) Not a date on which I am confident.
37. The impression I got clearly when the Secretary
of State came to the House two months ago was that there were
some details to be tidied up. I may be wrong, but I left thinking
that we are now 95 per cent of the way there and it is sensible
not to take the risk; and even if it was a 95 per cent risk, in
these circumstances I think that decision was right and, as I
have said, I do not think the Committee would take any different
view about thatin terms of the appropriateness of the decision.
Listening to youalthough it has been very helpful and I
do understand that you are being as helpful as you can and it
is useful to get all this backgroundthe project is receding
over the horizon as the session progresses. If you had asked meand
I am not a betting manI would have said there was a fighting
chance of getting this up and running by April 2003, and that
would have been bad news for the system but that would have been
possible. Listening to this, however, I think 2003 is optimistic.
The worry about that is that Ministers (because Ministers are
Ministers and they have got a political timetable which pays no
attention to IT schedules or testing or anything else) will want
this put in place and to be demonstrated to be working and demonstrating
benefits before the year 2005. I am beginning to get very nervous
now that we are going to get to 2005 and you might not have the
benefit of this system up and running. Is there any chance of
getting this in place by April 2003?
(Mr Smith) I do not want to get drawn into speculating
on when we are going to go live. Indeed, Ministers would not wish
me to do that because Alastair Darling has made it quite clear
that he wishes to make a statement to the House when he is satisfied
that testing has progressed sufficient to permit him to form a
firm conclusion on when implementation will be. However, I will
say that if our comments have induced that degree of pessimism,
we have not intended to induce that degree of pessimism. I think
the sort of dates you are suggesting are a little on the pessimistic
38. When do you think you will get from EDS
their anticipated "we will have finished testing" date,
which you can then assess as to its realism or not?
(Mr Smith) I think, in a sense, we have an on-going
series of discussions with EDS on what their anticipated testing
completion dates might be. When I will feel confident in their
anticipated end-date is when a number of the issues that they
are currently grappling with are showing clear signs of resolution.
We are at the stage, at the moment, where we are confident that
they know how they will crack these issues. Vince has indicated
that the technical experts, who have not touched the system before,
have reviewed those and confirmed that degree of confidence. I
am, by nature (going back to previous comments) something of a
pessimist. When I can see this happening and can see concrete
progress on these, as opposed to planned progress, then I will
start to re-think end-dates. I hope within the next month or two
we will start to move to a date where we can say to the Secretary
of State "Look, we are becoming increasingly firm about the
end-date of the system testing. You may now be able to go to the
House and make a further statement."
39. That answer suggests to me that, in fact,
EDS has already given you one or more anticipated completion of
testing dates and that you have said to EDS "We do not think
those dates are realistic." Hence your reference to "on-going
(Mr Smith) Going back a little while, EDS were telling
us not too long ago that they anticipated the testing date was
in advance of the April go-live date. We are entitled at this
stage to regard the conversation with EDS as being no more than
conversation, and these dates are discussion items rather than
a firm plan item.