POL00028390
POL00028390
Post Office Counters Ltd
Horizon Progromme Diector
Peter Crahan
Caps and Cards Business Delivery Director
Room 9
Holborn House
Caxton Road 30 June 1998
Eastway
_ Fulwood
PRESON PR2 4ZB
Dear Peter
Following the recent FMR meeting and the follow-up session in Newcastle last week,
my colleagues and I have continued to develop the options before us. Whilst
recognising that further detailed planning work is still required, we have become
increasingly convinced that in principle Option 2, as we named it, presents the best
way forward at this time. Our thinking on this is set out in the attached Paper. I
therefore thought it advisable to give you sight as early: as Possible of our views and
to do this before our meeting tomorrow.
Ihave also discussed the position at senior levels within Pathway and they are also
of the clear view that this Option is their preferred way forward.
We recognise the concerns you have regarding Pathway delivery this year and this is
incumbent upon us to mitigate these risks as much as possible. However, as
explained in the paper, there are advantages to BA and CAPS as well in this Option,
and a coherent position on fallback and contingency.
I would like to.make this a key agenda item for tomorrow afternoon’s meeting.
Clearly we need to understand the option to be addressed if we are to comprehend
the commercial and acceptance issues which it involves.
Post Office Counters Lid
King Edward Building
King Edward Street
London
‘ECT IAA
Telephone O171 776.3561.
roel
Krglanwd Save Lantos CAAA
Dwm186
POL00028390
POL00028390
-2-
Peter Crahan
30 June 1998
I would suggest that the agenda for 12.30 on Wednesday should be as follows:
1. Discussion of assessment set out in the attached paper
2. Decision on way ahead .
3. Discussion of acceptance and commercial issues around the chosen option
4, Definition of further work required to deal with these issues.
Yours sincerely
DAVE MILLER ec. an ees
fouh Woolhn
Sunn Alot
Dwm186
ASSESSMENT OF REPLAN OPTIONS
Background
At the FMR meeting on 18" June 1998 it was confirmed that the Benefits Agency could not support
the previously agreed durations for Model Office and End-to-End testing in the Testing Strategy given
the risks they now perceived in the assumptions that were being used in estimating these activities.
They proposed a revised approach (Option 1) which would result in the completion of Model Office
testing at the end of March 1999 compared to the previous target date of mid-December 1998. The
Horizon Programme was not content with the delay this would cause to the Live Trial and Rollout and
proposed, in outline, an alternative approach (Option 2). This split Model Office testing into 2 stages :
*. a phase from August to December 1998 which would test NR2 against CAPS 3.0 with Child
Benefit only, enabling a single benefit Trial from February 1999
© a phase from approximately January/February to May/June which would test NR2 against multi-
benefits, followed by rollout.
Whilst recognising that this option had significant contractual/commercial implications it was agreed
that Horizon and BA/CAPS staff should explore the 2 options further. This was done at a mecting in
Newcastle on 24th June 1998. This paper presents the Horizon assessment following these mectings.
Structure of the Paper
The paper is divided into three sections. Section 1 considers five key areas pertinent to the decision on
the two options, namely :
« Different Approaches to Model Office
¢ A Balanced View of Risk
e Fallback and Contingency
© The Contractual Position
¢ The Treasury Review
Section 2 reviews the two options in terms of the perceived advantages, disadvantages and risks to BA.
and to POCL. Section 3 provides further detail of the key planning steps and dates of each option.
Section 1 - Key Decision Factors
1. Different Model Office Approaches
The debate that has recently emerged around the replan is based on the different views and
expectations of the parties regarding Model Office management and execution. In summary BA/CAPS
believe that on the basis of their experience this testing will take longer than the Horizon Programme
estimates (primarily on the judgement of the required ratio of logical to physical days). Horizon’s
response is that their plans for Model Office cover limited objectives in that they do not expect to
address significant functional testing as this will already have been done in the various parallel streams
of testing that are being carried out at present (e.g. DIT, BIT, and Systems Test, as well as the support
to the CAPS JSA Model Office). Model Office is primarily a procedural test and the low ratio of
logical to physical days reflects the objective to test the system in ‘real-time’.
There are clearly difficulties in getting both sides to agree on this issue. This is one major reason why
Option 2 is proposed because it ‘decouples’ to a significant extent the first Model Office testing of
NR2 and allows it to be managed and executed more independently of CAPS. Also by using an already
proven CAPS version, and restricting the scope to Child Benefit, it reduces the overall uncertainty and
likely duration of testing. Option 1 on the other hand requires both parties to undertake joint Model
Office testing from August to March 1999 with a risk of conflict over scope, processes and objectives.
Pathway have already spent the last 6 months undertaking detailed systems and interface tests,
including joint tests with CAPS. They have also provided support to the JSA Model Office in the last 3
weeks without significant incidents so far. If the Model Office phase is primarily a procedural test,
(which all parties agree to in principle), then to spend 8 months testing in this area as proposed under
Option 1 seems generous in a programme that is already two years late.
POL00028390
POL00028390
2. A Balanced View of Risk
It is clear that BA/CAPS prefer a low risk plan, based on their experience, and one which will provide
a stable basis on which to plan the development and deployment with the feeder benefit systems. It is
also understandable that, on the basis of previous experience with Release 1c, they have doubts about
Pathway’s ability to deliver an ‘aggressive’ Model Office/End-to-End plan by Christmas 1998.
However, the more one raises concerns regarding Pathway’s delivery the less attractive Option 1
becomes. Although Option 1 allows more time for testing it nevertheless makes the BA/CAPS plan
dependent on Pathway keeping to these anticipated timescales in joint testing over the next cight to
nine months. If they do not then the CAPS plan automatically moves back as well.
Under Option 2, however, Pathway have at least five more months to test NR2 before they enter the
multi-benefit Model Office with CAPS. Only then is a tight alignment with the CAPS testing schedule
required.
Although an implication of Option 2 may be that CAPS do not complete all their benefit Model Offices
until 2 months later in 1999, the BA would in fact be in a position to rollout a larger number of cards
by the end of the year because of the earlier potential rollout to post offices under this Option.
The critical decision that therefore faces both BA/CAPS and Horizon is do they want to de-risk the
impact of Pathway late delivery by either :
a) providing a longer but fixed period of joint testing (as under Option 1); or
b) providing a more progressive two phase approach which allows Pathway to prove (or not prove)
delivery prior to commitment to a joint fixed schedule (Option 2) ?
The advantage of Option 1 is that it reflects the currently preferred CAPS schedules, but for this reason
creates a tight dependency that these will be jointly adhered to from August 1998 to March 1999.
The advantage of Option 2 is that if Pathway do deliver NR2 on time it enables an earlier entry to Live
Trial and rollout, protects Horizon and CAPS more in the event of delays by cither party, and reduces
the complexity of joint Model Office management and execution.
3. Fallback and Contingency
It should be noted that under Option 1 all contingency is built into the planned cycle durations. There
is no additional ‘spare time’ at the end of the cycles to absorb delays. If serious problems occur the
whole plan for both CAPS and Horizon will ‘move to the right’.
Under Option 2, however, if the first single benefit phase suffers delay there would be a potential
fallback of cancelling the planned release authorisation for Live Trial by Christmas and joining up with
the planned integrated JSA Model Office in January. Whatever the problems that may have occurred
under this scenario NR2 would still be in a better state, having had five months testing, to enter a joint
JSA Model Office than under Option 1. The penalty would be on Horizon not being able to start Live
Trial as hoped and having to fallback on the Option 1 dates.
Discussions on Option 2 to date propose that the second stage of multi-benefit testing should start in
January. Depending on what planning assumptions are used it will last until late April or early June. If
the sponsors chose it would be possible to start this stage later in order to provide additional genuine
contingency for the December delivery. It may be considered preferable to provide the contingency at
this point rather than allow a generous duration for multi-benefit testing, which will involve limited
additional functionality and procedures over and above what will already have been covered by this
date,
POL00028390
POL00028390
= _
4. Contractual Position
The work on the replan has not yet discussed in detail the ramifications of the two options on the
contractual and commercial position. Given the current ‘obsolescence’ of some of the contractual
conditions, it is expected that a serious renegotiation will be required in any event. This also appears to
be the position of the Treasury Review. Option 2 will require a more flexible approach by the sponsors
because it implies a more staged ‘acceptance’ process. It does however give more scope to reach a
revised contract with Pathway, given the ‘delay’ they may claim is being imposed on them under
Option 1.
5. Treasury Review
The Treasury Review has not yet finally reported and the timetable for Government approval of the
recommendations is not yet confirmed. Given that we anticipate that a consequence of a go-forward
option could be a serious renegotiation with the service provider, and a revised long-term strategy for
benefit payments, it is likely that the process will continue into the autumn. In these circumstances a
decision to enter into a closely coupled testing strategy from August may put additional pressure on the
parties at a delicate time. Option 2 however would give ‘space’ for Pathway and Horizon to complete
the testing of NR2 against the current CAPS Child Benefit platform while still enabling the carliest
practical rollout in 1999.
Section 2 - Perspectives on the Options
The planning dates show a difference in start dates between the two options for both Live Trial and
Rollout of approximately three months. Whilst the earlier dates are attractive to POCL, and Pathway,
it is not the only criterion on which a decision should be based. The degree of risk and the wider
impact of these options on the two parties also need to be considered.
1. BA/CAPS Perspective
Advantages of Option 1
The BA/CAPS view of Option 1 is that it best fits with their experience of their Model Office testing
and with their latest planned deployment of benefits in 1999. In particular it would allow them to
complete their Model, Office testing for all benefits by July 1999, rather than September 1999 as under
Option 2. However the difference in dates for the entry of a second benefit into Trial between the two
options is only two weeks (mid as compared to end June).
‘The risks they see in Option 2 may be summed up as follows:
their lack of confidence that Pathway/Horizon could complete their testing of NR2 and Child
Benefit by the proposed date of 18" December 1998
in particular, the ratio of physical to logical days.appears low to CAPS (1.5 to 1 under'the current
testing schedule for Model Office) -
also they believe that 2 runs of End-to-End testing may prove inadequate and 3 would be preferred
‘The section above on Different Approaches to Model Office seeks to place these risks in context and
suggests that previous CAPS experience is not necessarily a guide as to how Option 2 may proceed.
Pathway and Horizon are actively secking to reduce these risks, for example by reviewing the required
number of days in the test cycle,
The disadvantages CAPS see in Option 2 are:
« the two month delay in the completion of their benefits Model Office testing and the consequent
impact this may have on the development schedules of the Feeder Benefit Systems
«the possible need to establish more parallel testing environments
the reduced opportunity for testing with Pathway/Horizon during 1998 when only Child Benefit is
included (although it is believed that urgent stops and payments could be included).
While these disadvantages are recognised they must be set against the wider benefits in terms of
planning and flexibility which Option 2 would give to both parties as described in the sections above.
POL00028390
POL00028390
POL00028390
POL00028390
a ET
2. Horizon Perspective
Advantages to Horizon/POCL
As already noted an advantage for POCL of Option 2 is the earlier start of Live Trial and rollout. It is
important to stress that it is not simply a matter of an ‘earlier date’ but the opportunity to prove as
soon as possible, and address where necessary, the overall deployment in areas including:
training
installation
migration
communication
staff reaction
customer reaction
Although this approach may require contractual renegotiation, Option 2 is significantly more attractive
to Pathway given not only the earlier revenue but also the ability to reduce the costs of additional delay
with their suppliers.
The risks for Horizon in Option 1 are:
© the Model Office testing of NR2 is fully integrated with a CAPS multi-benefit Model Office from
day 1 giving reduced flexibility and opportunity to manage issues in the Pathway/Horizonw/POCL
domain.
the parameters which underpin the CAPS planning durations (while valid in terms of their own
experience) may not in fact apply to the new joint environment, whether on the positive or negative
side (a risk which is also relevant to CAPS)
© the differences between CAPS bringing a more stable code set to the start of Model Office testing
(given they would have already completed CAPS 3.0 and Child Benefit) and Horizon/Pathway
bringing the ‘first version’ of NR2 is likely to put the onus on problem solving on the latter and
result in tensions which could be avoided if the NR2 Model Office could be proven first in a more
limited environment
Pathway’s expected resistance to the delayed timescales that would be imposed by Option 1 is a
significant problem for its agreement that must also be recognised at this stage.
The disadvantages to Horizon of Option 1 are the converse of the points made above in terms of the
delay in entering Live Trial and rollout and the negative implications in terms of the relationship with
Pathway. It forces a close alignment of Model Office planning and execution over the next 9 months
and makes it virtually impossible to aim for an earlier rollout date than late September.
Section 3 - Overview of the Two Plans
1, Option 1
A copy of the high-level CAPS development plan showing Option lis attached at Appendix A. This
proposes a joint Model Office and End-to-End test with Child Benefit and JSA with the following key
features:
© start date -10" August 1998
duration - 33 weeks (including Christmas)
end date - 22™ March 1999
migration period - 6 weeks
Live Trial start date - 4* May 1999
second benefit into Live Trial (JSA) - mid-June 1999 (4 weeks after start of Trial)
completion of CAPS testing programme for IS and PSCS - 19* July 1999
If it is assumed that BA would agree to a Live Trial of 4 months, followed by a 2 week acceptance
period, this would result in a start of roll-out of Child Benefit and JSA in late September 1999.
2. Option 2
The high level plan for Option 2 is attached at Appendix B. As noted above this involves 2 phases:
2.1 Single Benefit Deployment - a Model Office and End-to-End test with Child Benefit only
(although it is possible that urgent stops and payments could be included in this scenario):
start date - 10" August 1998
* duration - 19 weeks
end date - 18" December 1998
migration period (including Christmas) - 7 weeks
¢ Live Trial start date - 8* February 1999
Again assuming a Live Trial of 4 months followed by a 2 week acceptance, this option would result in
a single benefit rollout (subject to contract renegotiation) starting by the end of June 1999.
A comparison of the Option 2 Single Benefit Deployment plan with Option 1, showing the more
detailed estimating parameters, is attached at Appendix C.
2.2 Multi-Benefit Deployment - a Model Office and End-to-End test with Child Benefit and JSA.
Different assumptions could be used for this test resulting in a ‘faster’ or ‘slower’ pace of execution .
Using this range of assumptions, as documented in Appendix 4, this Model Office/End-to-End test
would have the following features:
«start date - 18" January 1999
© duration - 14 to 21 weeks
© end date - 23" April to 11th June 1999
second benefit into ‘Trial’ - 10% May to 28" June 1999 (following a 2 weck acceptance of the
multi-benefit Model Office)
© completion of CAPS testing programme for IS and PSCS - by 11" September 1999.
As noted above sponsors may choose to plan a later start to this phase in order to provide additional
usable contingency at the end of the first phase. If an additional month’s contingency was provided,
and the multi-bencfit test was planned on a mid- -point between the faster and slower paces, then the end
date would still be around end June.
Under this Option sponsors may also not wish to commence the rollout of single benefit until the
multi-benefit Model Office had been completed and accepted, in order to give confidence in this
functionality. This view may still support a rollout date at the end of June. However POCL may choose
to rollout the infrastructure with a single benefit as soon as possible because of the other advantages
+ this would bring to them.
Under this Option roll-out of multi-benefit and additional cards would commence end October given
a4 month Trial period. The roll-out of multi-benefit would therefore start a month later than under
Option 1 (when Child Benefit and JSA rollout starts end September). However the rollout of single
benefit to Post Offices would have been taking place for 4 months already (July to October), providing
a much larger installed base for card roll-out.
For example, using conservative assumptions on beat rate of 200 offices a week and a weeks
“firebreak’ every 2 months, then by Christmas 1999 under Option 1, 2700 offices would have been
installed (300 in Live Trial plus say 12 weeks at 200 per week)
© under Option 2, 4500 offices would have been installed (300 in Live Trial plus say 21 weeks at 200
per week).
Therefore although Option 2 is likely to cause some delay in the completion of card payment testing in
CAPS, it offers the potential of a wider card rollout during the course of 1999.
Horizon Release Management
29* June 1998
POL00028390
POL00028390
POL00028390
POL00028390
bi
en
Major Strand, 1999
Sept I_ape_I May I Jun I tut_I Aug I Sept_I oc
se 22.ar
End o End Testing Runt +
End ond Tering Exe
yu ple Tremont
ISA Mode otice nant} ran f Fund SE
TauroctMo
hase ays) weet nad nee
estan Rant} Ran? Rand
TAUPO MedsOtiee I DARFOCLMead Othe TT BA cay MOTCARS5)
I
26Ape arjfon 19a
FSCS Mode Office Rant Ran2 4 Rand
TSCS Meda Ofe CAPRIS)
DLA/ WP Model Office / LCT
94a lense
Regression Testing t +
GR BAB ASG Regeron Tong
yal tie.
Integration Testing
Tihs € CAPS AO ntogaton Tong
29Mae 26Apr
POUppade
hse 2889 sae 10aty frm atsen) ins] ttang [689 408 25.00
Bulk Loads + ih
[Stand teas BAGont Posie owt [Powdered Pore] Pom I Pare [Pere? Fest Foe
May a1May
CCAS Solar Releases +
Nle20 CAPSS
Fr alle tw
eshva edkcsstne + t
once BACH BCRPSCr.
Statue: For Discussion
Dated: 19/6/98
POL00028390
POL00028390
Re-plan Option 2 Horizon Programme
Y Cray
2000
jer jevelea IS. I) Ot NASD: I d [oR I Mal Acs M™ lg Jo[b dele A, (Ss IeOnjeNe esp: I eset]
i ee eater ee
; : NR2:Model Office
i GhB / CAPS3.0}
i NR2 Release
> ChB: i
Start rollout
Second benefit I ?
Zinto Trial
Joini online :
payments/stopsand =
temp token tebting
S NR2+7/CAPS4.0
Multi-Benefit Model Office E
NR2/ CAPS3.0 / JSA / Is /P.
JSA Model Office
[CAPS3.0 /NR2
(PDR/CPR)]
I
I 22 June 1998 Page I
I
APPENDIX C - COMPARISON OF TESTING PROPOSALS
© Option 2 - testing of NR2 plus CHB only
© Option 1 - fully integrated testing of NR2 with CHB, JSA, IS, and PSCS
Activity
‘Joint? Model Office :
Start date
No. of cycles
Cycle length
Ratio of physical to logical days
Cycle duration
Reset period between cycles
Allowance for Christmas
Targetted testing/final fixes
Total clapsed time
End to End Testing
Start date
No. of cycles
Cycle length
Ratio of physical to logical days
Cycle duration
Reset period between cycles
Allowance for Christmas
Targetted testing/final fixes
Total elapsed time
Migration starts (5 week period)
Live Trial starts
Live Trial ends (est. 16 weeks) ,
Second benefit available
Start rollout (4 weeks after
Trial)
option 2
10/8/98
3
22 days
15
35 days (5 weeks) x3
Tweek
NIA
2 weeks
19 weeks (18/12/98)
10/8/98
2
33 days
2.5 (cycle 1) / 1.3 (cycle 2)
12 wks (cycle 1) + 6 wks (2)
1 week
NIA
O weeks
19 weeks (18/12/98)
41199
8/2/99 (ChB)
30/5/99
28/6/99
v7/99
option 1
10/8/98
3
24 days
25
63 days (9 weeks) x3
1 week
2 weeks
2 weeks
33 weeks (22/3/99)
10/8/98
3
33 days
2.0
63 days (9 weeks) x 3
1 week
2 weeks
2 weeks
33 weeks (22/3/99)
2913/99
3/5/99 (ChB)
20/8/99
14/6199
20/09/9F
POL00028390
POL00028390
I
APPENDIX D - OPTION 2, STAGE 2 MULTI-BENEFIT TESTING
The analysis below takes 2 views of this stage - a ‘faster pace’ based on more
optimistic parameters and a ‘slower pace’ based on the ‘standard’ CAPS planning
parameters. A faster pace may be possible given the testing that would already have
been performed on JSA and NR2 under Option 2.
Activity
‘Joint? Model Office :
Start date
No. of cycles
Cycle length
Ratio of physical to logical days
Cycle duration
Reset period between cycles
Allowance for Christmas
Targetted testing/final fixes
Total elapsed time
End to End Testing
Start date
No. of cycles
Cycle length
Ratio of physical to logical days
Cycle duration
Reset period between cycles
Allowance for Christmas
Targeted testing/final fixes
Total elapsed time
Acceptance to enter Multi-
benefit Trial
Multi-Benefit Trial (16
weeks)
Multi-benefit Trial starts
Multi-benefit Trial 2 ends
Start Multi-benefit roll-out
‘Faster pace’
18/1/99
3
15 days
15
24 days (3.5 weeks) x3
1 week
NA
2 weeks
14 weeks (23/4/99)
18/1/99
2
33 days
1,5 (cycle 1) / 1.3 (cycle 2)
7 wks (cycle 1) + 6 wks (2)
1 week
NA
O weeks
14 weeks (23/4/99)
2 weeks (7/5/99)
10/5/99
2718199
31/8/99
‘Slower pace’
18/1/99
3
15 days
2.5
37 days (5.5 weeks) x 3
1 week
NIA
2 weeks
20 weeks (4/6/99)
18/1/99
2
33 days
2.0
63 days (9 weeks) x 2
1 week
NIA
2 weeks
21 weeks (11/6/99)
2 weeks (25/6/99)
28/6/99
15/10/99
18/10/99
POL00028390
POL00028390