Examination of Witnesses (Questions 500-519)
MR SEAN
SHINE, MR
PETER HOLMES
AND MR
ANDY NAISH
22 MAY 2006
Q500 Chairman: Your end at the bit
that says, "This is how much you ought to receive".
Mr Naish: I am not quite there
yet. Once we have established the entitlement we calculate the
value of that entitlement and then we create a record that indicates
the payment to be made for that claim, the amount of money that
will be paid for that claim. That system then creates a task for
a user to perform to authorise that the payment be made. It delivers
a task to a user on a screen which says, "Please authorise
this set of payments together"; it creates a batch of payments
that are authorised together. Once a user says, "Yes, this
payment is authorised" that night our system will then create
a file of payments which then go into the payment system itself
which executes the physical payments.
Q501 Chairman: Let us go upstream.
You were saying that after validation you began the process of
entitlement calculation.
Mr Naish: Yes.
Q502 Chairman: Which bit of validation
were you responsible for, if any?
Mr Naish: The system validates
the claims in two steps. First of all a set of checks that the
claim form itself is correctly completely, that the customer exists,
that the details of the customer match the records. There is then
a second stage of validation that checks that the land associated
with the claim is matching the land as recorded in the Land Register
to make sure the land is accurately being claimed. Our system
validates those things and creates tasks for the users if there
are any issues associated with that validation, any points of
discrepancy.
Q503 Chairman: Do you do anything
before validation?
Mr Naish: Before validation we
do the claim capture process.
Q504 Chairman: You do from claim
capture to putting a message on a screen that says, "This
is how much the farmer should receive". Your system is responsible
for everything between those points of activity.
Mr Naish: Yes. There is a wider
business process where users are making use of our system to clear
validation issues that have been raised.
Q505 Chairman: I am trying to establish
very clearly what you are responsible for. I want to move into
this bit and try and unpack what, if anything, did not work properly.
We are getting somewhere near to being able to try to disaggregate
this system you have created to see what did not work. What we
do know is that there were huge problems with the processes involved
in validating claims. The Committee has heard evidence on this.
You have said to us that all of the system worked. You told us
the Rural Land Register went live in September 2004 so I presume
that meant it worked and you are satisfied it worked. You told
us the Customer Register went live in February 2005. The High
Volume Data Capture went live in May 2005. The core validations
functions for the single payment went live in a pilot in July
2005 and rolled out in August 2005. If all of that worked why
were there such problems when it came to do it? I think that is
what Sir Peter was after. You have told us you delivered what
you were asked to do; you have produced a list for us of the things
that you say worked, and yet when it came to claims validation
the thing collapses. Why?
Mr Naish: I do not recognise the
collapsing.
Q506 Chairman: We know there were
a lot of problems with claim validations. Ministers told us in
relation to the number of applications, when they finally pulled
the plug on March payments, it was an indication that there were
a lot of claims that had not been fully validated.
Mr Naish: Yes.
Q507 Chairman: So we know the validation
process had some problems. Why?
Mr Naish: The primary driver is
one of volume. The amount of work, the number of tasks that were
created by the system where there were discrepancies in land from
claims that were on the Rural Land Register, was much higher than
had been expected.
Q508 Chairman: In paragraph 27 of
your evidence you told us that the January 2003 contract would
be capable of supporting 100,000 customers was increased to 150,000
customers as part of the revised May 2004 contract. That would
suggest to me that you anticipated extra business going through
the system. If there was more than 50-odd thousand customers going
in by definition there must have been more bits of land going
to be applied for. When did you first become aware of the volume
issue? Who told you or did you just wake up one morning and think,
"Oh, good heavens, we've got all these extra things to deal
with. We thought we'd got it all bolted down, everything was working,
everything was fantastic"? What volume of land were you working
on?
Mr Shine: If I could comment on
that, I think I have covered some of it already. The expected
volume of land changes, as I said, was predicted beforehand to
be in the region of around 10,000 per year.
Q509 David Taylor: Predicted by?
Mr Shine: By RPA as part of our
contract. As you would expect as part of our contract we will
design a system for a particular volume so one of our pre-contract
assumptions in terms of procurement time was to take some view
on what the volume would be. As I have already said, the actual
volume of land changes turned out to be much higher than anyone
had predicted, in excess of eleven times higher. Let me take a
real life example in terms of understanding what might happen.
There may be two farmers/land owners who both claim on the same
piece of land. It may be that the boundaries between them have
moved somewhat; it may be that one of them has drawn the map incorrectly,
but that one claimed for one plot of land on one side and one
plot on another. If there is an overlap there is a task that says,
"The application that you have made for a piece of land does
not hang together correctly because those two boundaries do not
work". That issue will in turn generate some tasks which
would therefore require somebody within RPA to look at that and
in some cases they might contact the farmer directly or in other
cases they might look back at the original and take a view from
there. The volume of tasks is a function of the large number of
land changes and indeed the shift of focus to land. The underlying
focus of CAP reform was to shift away from production to that
of land, therefore the number of land changes is much higher,
therefore the number of tasks generated in terms of validating
that particular claim was high.
Q510 Chairman: On a point of detail,
does that mean that the original estimate for the volume of land
that you might have to be dealing with, was that predicated on
the historic system?
Mr Shine: That is correct.
Q511 Sir Peter Soulsby: Can I follow
up this question about the volume of claims and ask Mr Shine when
they first became aware that the volume was not really what had
originally been proposed and was in fact many times more?
Mr Shine: Certainly from recollection
that seemed to occur round the July/August 2005 time period. It
was during that period because that is when the amount of work
on the system increased, the amount of changes coming in, so it
is in or around that time.
Q512 Sir Peter Soulsby: You are telling
us that one of the main contributors to the difficulties that
subsequently emerged was the increase in volume. You had been
aware of that increase in volume back many months earlier, but
nonetheless you were taken by surprise when the whole thing failed
to produce.
Mr Shine: No, what I said is that
the increase in volume in land changes did result in some technical
issues with regard to system availability which were dealt with
immediately at that point. Essentially we changed the system and
as you are also aware the RPA took some steps to outsource some
of the digitisation of some of the land changes that were coming
in as well in order to give increased capacity in a short period
of time. That was the case then. As I have said already, the IT
system component then required to enable the entitlements to be
made and to enable the payments to be made was there and running
from that period.
Q513 Sir Peter Soulsby: I understood
you to be telling us that it was this increase in volume that
was the major contributor to the difficulties that subsequently
led to this project failing to deliver. If that is the case, how
come, if you had been aware for many months before that, you were
taken by surprise when it did actually fail to deliver?
Mr Shine: The increase in volume
was a contributor to the increased number of tasks which subsequently
had to be worked through by the RPA.
Q514 Sir Peter Soulsby: Surely you,
as the ones who are intimately involved in developing these systems
and are responsible for these systems, must have realised at a
very early stage that the volume that was actually going to have
to be dealt with by the system was out of all proportion to what
was originally expected. Surely you must have realised that that
would have implications on whether or not the project could actually
be delivered on time. Yet, you tell us, despite having had this
knowledge you were nonetheless taken by surprise when it failed
to deliver.
Mr Shine: And very much our focus,
you are right. At the time when the amount of land changes were
higher we had to do immediate work in terms of increasing the
capacity of the system to handle that volume of changes which
we focused on during that period and which we did. What we were
not managing and were not responsible for was actually managing
the business process. In effect the number of tasks that were
being handled and how those tasks could be closed and how fast
those tasks were being dealt with, that is part of the business
process.
Q515 Chairman: Who is responsible
for that?
Mr Shine: That is the RPA.
Q516 Mr Drew: Can I establish very
clearly in my own mind, these are bespoke bits of software; they
were written especially for the RPA initially to take account
of the historic methods of payment but you were then given separate
enhancements to go on and adapt the software. Are these very complicated
bits of software? On the A to Z of software creation are we on
the As or are we on the Zs?
Mr Naish: It is difficult to make
a comparison but they are very complex calculations to work out
entitlements and payment amounts. I think the key area of complexity
is in dealing with the Rural Land Register and the mapping process
around that. It is scale which drives complexity and the way in
which land in particular is handled makes it complex.
Q517 Mr Drew: As soon as the system
approaches change and the RPA came to you and said, "We are
not going for a historic system; we are going for a different
system" you must have said, "That's fine, we can do
this but don't come to us telling us there is going to be a big
increase in the volume of people putting in claims". That
is what you would have said.
Mr Holmes: What we said actually
was, "Let's sit down and understand the requirements and
have discussions about it, which is what we did.
Q518 Mr Drew: At what stage did you
go from discussion to having some concern that if there was an
increase in the volumes—whether it is the volume of the
land numbers or the volume of the claims—to start saying
to the RPA, "Volumes are going to make a difference here"?
Mr Holmes: We went from discussion
to agreeing the revised contract in May 2004. We then had separate
work going forward about the complexity of the new policy. As
we have said a number of times today it was only later in the
year that some of those detailed requirements were finally understood
and a system was sized to meet the volume of customers that we
ended up dealing with and that was fine. I guess we started to
understand the volumes involved in terms of the land changes around
the spring of 2005. In fact we made some changes to the software
to handle high volume data capture. We were taking on board the
IT system implications of the increased volumes. We produced some
software to allow the data to be captured at volume and we did
some work to cope with the increases, but we were not looking
at this from a "how will all these forms be processed, how
will all these tasks around the land changes be handled"
because that was outside the scope of our contract.
Q519 David Taylor: You have just
heard your colleague say that scale drives complexity. Do you
agree with that?
Mr Holmes: Yes.
| 
