Examination of Witnesses (Questions 540-559)
MR SEAN
SHINE, MR
PETER HOLMES
AND MR
ANDY NAISH
22 MAY 2006
Q540 Chairman: That was data which
had come from the mapping exercise.
Mr Shine: Yes.
Q541 Chairman: The Committee has
seen many, many cases where when farmers sent in their details
and it came to validating their claims, there seemed to be a lack
of fit between what the farmer said was their parcel of land and
what the Rural Land Register said belonged to them. Then all kinds
of strange things happened. Land was uplifted and put somewhere
else, all kinds of operational glitches occurred. What I am coming
back to, bearing in mind this information that farmers had was
available at the same time that you had satisfied yourselves in
September 2004 that your Rural Land Register software worked,
did nobody actually go and get some examples of what farmers said
was their land and see if it worked in the Register before you
actually went live with it?
Mr Shine: My understanding is
that from that period onwards there was a continuous amount of
land changes happening. As I have said already, the standard practice
is that there would be changes . . .
Q542 Chairman: Mr Shine, that is
not what I was asking about. What I envisaged was that somebody
might have gone out to a group of farmers and said, "Excuse
me gentlemen, would you be so kind as to help us? We are going
to test out the Rural Land Registry. Would you please submit us
in the following format what you think your holding is?"
You would have then said, "Look, this is the front end of
the application, let's run it through the Land Register and see
if everything fits together using the farmers' supplied rural
information". Did that process not happen until you went
live with the system?
Mr Naish: I cannot say whether
it did or not. I could perhaps be allowed to go and find the answer
to that question.
Q543 Chairman: Mr Naish, I cannot
believe that you do not know whether in fact your piece of software
which you believe was capable of delivering was not in any way
evaluated against a real world land data from real farmers until
you went live.
Mr Shine: Perhaps I should clarify
this.
Q544 Chairman: That is a good idea,
Mr Shine.
Mr Shine: The 500,000 parcels
that were put in September 2004 was real live land data. It was
based on the digitisation process that had been undertaken by
the RPA with real land.
Q545 Chairman: The point is that
there is a difference between building the land database, what
you think is out there based on these satellite pictures, ordnance
survey and co-ordinates and all of that going in . . .
Mr Shine: And, if I may say, land
submissions from farmers as well. It was not just a theoretical
exercise.
Q546 Chairman: If that was testing,
why did it go so wrong when it came to go live? Why have we heard
so many complaints about the mapping system and that seems to
be at part of the heart of why validation could not take place.
Why did that go wrong?
Mr Shine: I think it is worthwhile
saying that first, as I have said, the shift in policy made a
profound movement from that of production to that of land. It
did come out in last week's discussion that while there was an
expectation within the RPA that the farmers would have had their
land records up to date from previous interactions they had had
with the Department, it turned out that that was not the case
and in fact that the volume of changes made by farmers and made
by land owners was much higher than expected.
Q547 Chairman: I accept that there
may have been in process timing a volume issue, but what we are
talking about is an accuracy issue. In other words, somewhere
along the way what the farmer sent in I presume did not tie up
with what was in the Rural Land Register. Is that right?
Mr Shine: That is correct.
Q548 Chairman: Is that because farmers
were wrong or they changed at the last minute? Why could it not
be sorted out?
Mr Shine: You depend on getting
input from the farmer. The scheme was based on the farmer saying,
"Here's my land" and giving mapping. Part of the process
that went on was that the RPA sent maps out to the farmers and
said, "Here's the reference that we have about your land".
Because of the shift in policy to move to land my sense is that
the farmers were very much focused on having a higher degree of
accuracy in their land because it also results in entitlements.
As you are aware the entitlement process was going to be done
once in terms of establishing definitive entitlements; that was
one of the features of the new CAP reform scheme. Therefore there
was an onus on all land owners to ensure that their land records
were fully up to date. That was one side of the process for existing
land owners. The second part of the process was that there was
now a new number of customers—as I have said already, over
40,000—who were now focused on adding more land details
because they had never had them in there before. Those two things
together resulted in a huge volume. You also then had the possible
situation—but real life situations do occur—where
two farmers, for example, on both sides of a land boundary perhaps
moving it a few feet away.
Q549 Chairman: In reality you, as
Accenture, only became aware of the changes you have just described
from the time when the system went live and started to deal with
real live applications.
Mr Shine: Yes, but secondly I
think it is worthwhile saying—as you have said already—that
the closing date for the scheme is when the large volume of changes
came in and that is when the business process said, "We now
have all of these changes to make, let us start entering them"
and that resulted in the system having much higher requirements
for throughput than what it was designed for.
Q550 Chairman: What I am intrigued
about is that you said you only became aware of this in July 2005
but the closing date was May.
Mr Shine: As I have said to you
already, my precise memory on the precise dates is not fully accurate.
I am very happy to check that but it was around that period.
Q551 Chairman: In your judgment,
looking at this, was there anything that the RPA could have done
to have trialled in the real world some of the implications of
the policy changes and given you a better feel for the volume
side of things? Let us be benevolent to you, you look as if you
are feeling in the dark; the customer is not giving you accurate
information from what you have told us about the anticipated volume.
The customer has sent out to all these farmers maps which do not
tie in with what the farmer believes is the land holding situation,
and when all of this lot comes hammering back with 10 times the
volume that was anticipated, not unnaturally something had to
give and I presume it was the inability of your system not because
from your judgment it was badly designed and incapable of delivering,
it just could not cope.
Mr Shine: Partially, but the bigger
impact was the volume of forms and all the validation checks and
tasks that had to be done.
Q552 Chairman: Although the closing
date is May 2005, if the RPA had been sampling the volumes coming
in and the content, I am sure they would have been able to predict
that something was happening that they did not expect. Did they
not communicate at all to you in April or early May and say, "Hey
guys, something is happening in the way these applications are
coming in that we didn't expect".
Mr Shine: As I think was said
last week, it is down to customer behaviour in terms of putting
in changes at the last minute. I cannot recall precisely the time
when the significantly higher land changes began to become clear.
I am happy to go and check that and come back to you in writing.
Q553 Chairman: I would be grateful
if you would because, as I said right at the outset of this inquiry,
we are not trying to stitch somebody up. We are trying to find
out what actually went wrong and who must accept responsibility.
We started off with one policy, the historic; we moved to the
dynamic hybrid which had certain consequences. You are telling
us that in volume terms from the Accenture standpoint that was
not predicted. Suddenly your customer confronts you with a lot
of information which you did not anticipate and your system which
was, according to you, fit for purpose could not cope. In terms
of actually putting all of this extra information in was it the
case that the RPA could have got round that by having more operators,
more input, or was the system only capable by definition of taking
in and dealing with so much information at a time?
Mr Shine: Again to clarify the
volumes question, as I have said already the system is designed
to support up to 150,000 customers, so the number of farmers,
the number of land owners up to 150,000. The system was also designed
with an expected volume of land changes as to what happened in
a period of time. As you said yourself, Chairman, it is the volume
of those changes in a period of time being 10 or eleven times
the expected volume which caused the temporary issue in terms
of the availability of the system which we identified and focused
on. We made some changes to both hardware and software in conjunction
with the RPA. We made those changes and the system is available
and does support that volume of changes today.
Q554 Lynne Jones: We have heard about
those problems but you seem to have overcome them because you
were not flagging up any problems to permanent secretaries or
to the ministers. Is that correct?
Mr Shine: As I have said already,
I recall the specific meeting in September 2005 where much of
the discussion was focusing around some of those system availability
issues.
Q555 Chairman: Who was that meeting
with?
Mr Shine: That was with the permanent
secretary.
Q556 Chairman: And who else?
Mr Shine: With representatives
and officials from the RPA and a number of people from Accenture,
including myself.
Chairman: We are going to suspend the
Committee for as short a time as possible, 10 minutes if colleagues
can manage it, whilst we go to vote and then we will come back.
The Committee suspended from 6.14pm to
6.26pm for a division in the House
Q557 Mr Drew: Can I reflect on what you
have been saying to us and I think I would like one question answered
from what you have been alleging. With the benefit of hindsight,
if you had been able to run a full scale test with real data that
could only have been a good thing. Or was what you were trying
to do so clearly identified in your minds as a company you just
thought it was going to work so a test was irrelevant?
Mr Holmes: No, we would always
say it is a good thing; it cannot be anything but a good thing.
Q558 Mr Drew: So why did you not
ask for real data to do a test? Why do you think the RPA did not
offer you real data to do a test?
Mr Holmes: Real data to do a test
at which stage?
Q559 Mr Drew: At a stage when you
had the data coming in, from May 2005 onwards.
Mr Naish: We did do a test on
real data. Once the final part of the end to end system had been
delivered and the processing of claims through the validation
steps had gone far enough in January or February this year, this
was the earliest opportunity to take real live claim data that
had been progressed far enough through the business process to
be able to test those entitlement and payment steps.
| 
