The end of Kiwibank?

<updated to clarify this is for Kiwibank’s Core Banking System>

Fairfax Media report that Kiwibank is signing up to SAP as their “principal computer platform” provider for their core banking system. The core banking system runs the business of the bank – including real time transactions, interest calculations and payments., and it is unclear where the line is drawn between this and the other systems in the bank. 

If the reports are true, then that’s arguably very problematic.

SAP, and I’ve been through a few cycles of SAP rollouts, is a vast, complex and barely usable system for managing complex businesses. It has a history of failure, but I also sense that much of the failure is unreported.

SAP and other ERP systems are sold, as I understand, on making it easy for the most senior managers to manage their sprawling businesses. But SAP and other systems like it generally make it very very difficult for the huge numbers of end users inside the company to actually do business. The recurring joke is that  SAP promises to save tens of millions, costs hundreds of millions and causes billions of dollars in lost productivity and opportunity.

I can’t back that up, as I have never done a study on it, and I could never talk about anything I have seen, but I can refer to the AU$1.3 billion and six years that Commonwealth Bank reportedly spent on upgrading <update – it was a new install for core banking, and they run Peoplesoft/Oracle for their ERP> their SAP system to be able to do live transactions and for compliance with new laws. Note that ASB, which CommBank owns, coded and delivered their own live transactions banking system themselves back in 1969 – and it still works. And also note that in September last year Tom Groenfeldt reported only two SAP core banking replacements – CommBank and Nationwide Building Society in the UK.

CommBank is huge, and arguably needed the complexity of something like SAP. While the project was a big risk, they were able to throw vast resources at it and no doubt customise it to better fit their business. It looks like they got away with it.

Moving to a system like SAP (or installing any large IT project) is a very risky endeavour, and generally means changing the way you do business. The waterfall-like approach to these projects means that requirements are set years before delivery, and the rigidity and complexity of software like SAP almost always means failure to deliver on promised costs or benefits.

For Kiwibank I worry that they are too small, with only 850,000 customers, and with little money to invest, they will have to use the out of the box solution. And to quote Groenfeldt on smaller banks “When they do a core replacement, they make a fundamental decision to adopt the structures and procedures defined in the vendor’s package with minimal customisation.

So I worry not just that the cost of implementation will be high and the result may be poor, but that the very nature of what Kiwibank is will irreversibly change.

But then again I know very little about SAP for Banking, and I hope for their customer’s sake that I am wrong.

The ideal solution is, of course, for Kiwibank to wake up to the very strong  local development talent, hire them in and give them true power and air cover to reinvent banking, piece by piece and digital-first.  It’s that approach is good enough for the entire UK government, it’s good enough for a tiny antipodean retail bank. 

Published by Lance Wiggs


41 replies on “The end of Kiwibank?”

  1. I do not know what your background is and where you are coming from but it is not the SAP system’s fault , it is just the wrong people that implement it by the customers. Many SAP Banking projects begin with “just local resources” with a hope that they will learn fast and everything will be OK. This wishful thinking leads usually to 1 to 2 year delay. After that the customer usually realizes that the experts that looked very expensive at the very beginning, do not look expensive after all. It is however too late, the money has been burnt out and the new people usually need to start over or clean the incredible mess that the cheap resources left after them. I have worked in 4 SAP Banking projects worldwide by now (Latin America and Europe) and the mistakes that had been made were all the same. People do not learn from their own mistakes.


    1. “it is not the SAP system’s fault, is just the wrong people that implement it by the customers.”

      Blaming a poor product on the customer ignores the central problem of the product itself.

      Your comment as an expert is doubly troubling as it also reinforced that SAP Core Banking projects are likely to have “customer mistakes” and be a lot more expensive and lengthy than first thought.


  2. As a past user of new rollouts of both Peoplesoft and SAP, I agree with Lances comments – We saw very major downsides for the organisations as wholes, with senior managers getting a possible upside though never confirmed to be as significant as advertised (and should we talk about the sponsored fact-finding trips around the world for the recommending teams?)


  3. I have observed, but not participated in, four major SAP implementations in organisations that were our clients at the time. One in Australia, two in Europe and one in the USA. In each case the timeline and the budget ran well over original estimates, by which time it was considered too late to change to a different platform. The outsourcing companies which are often hired in to do the implementation, usually make very large amounts of money from the customer, whose staff are left to deal with something very clunky. The prospect of making further changes when requirements inevitably have to be updated is very frightening.
    In banking today, everything is mobile and digital, and changing with incredible speed. Not only must the platform be ‘natively digital’ but it must also be capable of very rapid evolution. I do not see SAP in this category at all.
    Very worrying for Kiwibank and all its stakeholders.
    I hope they will be okay.


  4. While I believe that an SAP implementation is going to be expensive and problematic, I also think this solution is probably the least worst option. The current software, Ultracs, is a piece of credit union software, and Kiwibank currently spends a fortune every year on custom development to keep this software fit for purpose. It is fundamentally unsuited for providing corporate or large business banking services.

    There are very few options for core banking platform, and the idea of having one developed from scratch is frankly ludicrous. Yes, ASB did it – forty five years ago, and it has evolved since then gradually. I note that it does cause them some issues, have you tried to have a password longer than 8 characters in their internet banking?


    1. ASB are currently migrating their core banking platform to SAP over a phased implementation. They went live back in early 2012 and are continuing to migrate business functions from their in-house developed mainframe solution to SAP.


  5. I was in the trademark replacement project at Asb as part of the 2billion system upgrade initiated by cba. The core banking replacement–use sap for multi currency support did not finish in time nor budget. I had to do alot of interfacing work to the sap banking module. When there is anything that can be done outside of sap, appropriate or not, it will be done outside of sap. The reason being its too expensive to do anything with it. The consultants were charging 5000 nzd per day per head and they are not real developers….


    1. And the fact that they are not real developers is bad because….?

      You buy SAP not to develop. You buy SAP to reuse 70 to 80 percent of what is delivered in so called standard. On the other hand if someone buy SAP to do exactly the opposite (70 to 80 percent), then he is crazy. If one does not understand that, then do not let him buy SAP.


      1. My personal experience on using SAP in the UK was working for Barclays Capital (where I managed a team developing custom solutions that interfaced into SAP) and it was certainly a successful implementation there. The business was extremely complex, operating in over 100 countries globally with over 26,000 staff. I think the key thing, as has been mentioned, is this the management, technical resources and financial resources were absolutely first class throughout the organisation so everything was implemented and built correctly from the ground up. In fact, dare I say it, a project I was involved with around global tax reporting was almost a pleasure to work on.

        Frankly, having been back in NZ for a while now and contracting (and in particular the experience at a previous contract I held at an SOE which was quite an eye-opener into how poor the management of enterprise IT here in NZ can be) the management quality to oversee a project like this will be the biggest risk.


    2. The consultants on the ASB project were not charging 5000 nzd per day … no where close to that. Yes, the go-live was moved, but the success of the project in early 2012 can not be refuted. The ASB SAP implementation continues to migrate functionality from the existing core-banking solution to SAP as part of its phased implementation over the next several years.


  6. You think that building an entire incredibly complex system from scratch, using developers whose prime experience is in smallish websites and in a country with some of the worst managers in the developed world is likely to be risk free? With all of it having to be working before cutover (or having incredibly complex interfaces to largely undocumented legacy systems). You can’t just go live with a Minimum Viable Product having no liquidity gap analysis, or SWIFT interface, or overseas credit card payments.

    Maybe they should have gone to a specialist banking system provider rather than SAP though. That I’d agree with.


    1. Actually, if you did want to build from scratch, the best plan would be to start a new bank Kiwibank2 without any customers and a closely specified target market, then as the system became available, start adding customers really slowly before migrating people from the old bank once it was completely stable and functional.

      Banking is a mass of corner cases – like a mortgage jointly to a trust and an individual, but with only one of the trustees liable for the debt. And if you don’t have every last function available online, people will sneer at you for living in the 20th century.


  7. I have worked on complex software projects including with a couple of the major vendors. What always happens is that the real project risks around these projects are not well understood or managed.

    Once the sales team are replaced by the delivery team it can be almost impossible to really track all of the deliverables in a meaningful way.

    Think of it as a proposition bet. The client thinks they are buying a static fixed outcome but the reality is always somewhat more tricky than anyone thinks. And – this is the key thing – the outcome is always different.

    Consequently a prospective buyer needs to do advanced due diligence on the product & services bundle and get absolutely full disclosure of everything. They also need to talk with risk managers at other clients ( like Commbank) who will probably not be allowed to talk to them.

    A site visit hosted by the vendor will be offered but those are like visits to North Korea where you won’t get to see or assess the real state of the affairs and more than likely will be fronted by pr groomed and coached people.

    Kiwibank needs to ask the hard questions and make devils advocate type scenarios such as if costs double & timeline increases by 300% what happens then. The failure points for these projects are known and they are spectacular.

    I read about a Queensland project involving both SAP & IBM that snowballed to the point where a $6m project is now a billion dollar problem and a huge lawsuit.

    They quote a McKinseys report which says “Half of IT projects with budgets of over $US15 million dollars run 45% over budget” which I’d say is on the low side.

    The NZ payroll project for education is a rerun of the Queensland project which was a payroll for health workers.

    In all complex software projects there are moving elements in the equation ( on both sides) and NZ managers and companies can very easily get crunched by international consultants who play hard ball with governments so a small bank in Wellington is nothing for them.

    I wonder whether a hosted bureau service where the client actually pays for a working system would limit the risks. Later if that works they pay extra to bring the service “in house” in a 2 stage project.

    That way the vendor has to prove everything works in a controlled environment.


  8. Hi Lance.
    I was at the Commerce Committee hearing where Sir Michael Cullen let slip that Kiwibank had chosen a bidder to replace Ultracs after a one year search had narrowed it down to 2 bidders. The announcement is due next Monday.

    Sir Michael told us afterwards Kiwibank had got a good deal to buy a system that was being used by one of the big four. He said most of the work to implement it would be done in-house.

    Question for you and your readers: Who was the other (losing) bidder? And why might they have lost?

    I suspect the Reserve Bank has a particular interest in all this from a systemic stability point of view.



  9. Lance

    Yes I share your concerns – they overlap with a piece I wrote last year for the NBR on the government procurement, (“everybody knows that most big software projects fail. It’s simple: the complexity is too big in relation to the rate of change, therefore unpredictability)”. I guess one has to assume that Kiwibank has some very smart people who have put considerable time in to evaluating options and speaking to others who have gone a similar way. I may be a little naive here (having never worked with SAP or core banking systems) but surely there must be ways to use a big, slow moving products like SAP but still allow agility/innovation around the edges (i.e. expose well architected API’s that allow developers who don’t know about SAP to do cool stuff).

    I suppose the problem will always come back to vendor lock-in. Once that x year contract has been signed and Kiwibank has started the project, their negotiating position deteriorates very quickly. Not a whole lot of incentive for SAP to cost effectively modify/extend their API to allow more stuff to be done outside their product a year down the track.

    Not sure what the answer is but glad I’m not making the decision. But as you say the UK government has proved that sometimes things can be done very differently – I love this article

    It would be great to think that a small innovative bank in a small innovative country could find a way to avoid the seemingly inevitable process that many large organisation get into of getting locked in to slow moving product, massive time/cost overruns and the loss of innovation/agility.

    The banking industry will be disrupted over the coming years and as a smart chap called Charles Darwin once said “It is not the strongest of the species that survive, or the most intelligent, but the one most responsive to change”.


  10. A friend of mine specialises in SAP implementations globally.

    Their day-rate would make most peoples’ eyes bleed, including people with international experience. At first I thought it was a 5-day rate on the low side.

    I gathered from our many discussions over 15+ years that SAP is a very, very well though-through product in terms of its profile, marketing, selling, consultancy and reputation. Quality, reliability and value, not so much. It is an impressive system ecosystem of salespeople, consultants and other actors. There is a lot of effort that goes into understanding the potential client, the pressures and demands upon them, and how to make the stakeholders feel comfortable for their own (personal) sake (career, accountability, embarrassment, etc.).

    For somebody who just wants to make oodles of cash, SAP consultancy appears to come highly recommended. For those with a conscience and/or belief that value must ultimately be delivered, another vocation is apparently recommended.


    1. “I gathered from our many discussions over 15+ years that SAP is a very, very well though-through product in terms of its profile, marketing, selling, consultancy and reputation. Quality, reliability and value, not so much.”

      And yet over those years there have been more and more customers that decide to buy and implement SAP in their organizations. Gosh, are they all hypnotized, or what?


      1. To be clear, I’m not saying that SAP never gives anyone any value.

        To your point, I would say that the thing about businesses and organisations, especially of size, is people very rarely admit that they didn’t understand/do their job properly/that the salespeople really got the better of them and ran away with the money.

        This dynamic is very well-known, and I think you know exactly what I’m talking about. Nobody admits they are incompetent, so everyone else assumes all is well. Then their lack of negative feedback turns into a “happy customer” for the next pitch.


  11. I’ll echo the majority of sentiments here where SAP is sold into corporates as the holy grail to managing core systems, products and customer interactions. The reality is a world of financial pain as implementation projects over-run on time and under-deliver on promises. Just ask Contact Energy….


  12. The other platform in the fram is likely to be Oracle Banking – as implemented by NAB.

    Funding the project is an issue not just for the Bank but for the tax payer. Kiwibank has yet to pay a dividend after 10 years of operation and all profits are being ploughed back into regulatory capital, core banking replacement and building it’s branch network as Post reduce their footprint and change the basis on which Kiwibank paid for access to their network. The often touted customer base of 850,000 needs to be treated with caution as it doesn’t necessarily mean active or profitable. Some of the banks’ profits are a result of moving payment services from Post to the Bank.

    Why the tax payer carries the burden of an bank that has arguably done little to change the banking landscape (other than to play on and build anti-Aussie sentiment) and is currently facing a class action lawsuit with respect to it’s excessive fees is beyond me. The same and bigger effects could have been achieved through regulation/regulatory bodies such as the commerce commission.

    Contrary to its advertising, I’m not sure Kiwibank has ever been revolutionary (and the current class action seems to indicate it was as mercenary as this big four in fees). So why would they do something revolutionary like build their own platform – would help support the Wellington “valley” and the NZ economy?


  13. I work for SAP. I am not authorised to speak for them here so these are my personal opinions / comments.
    SAP is not implemented the way you install Flappy Face or Magic Birds. To get it live (properly) will take experts. If you think you can slap it in and it will work you are unfortunately wrong. And if you think the guys that can put it in properly are expensive, check what fixing a mess will cost you and then think again.
    That aside, just a comment. Why would SAP be making double digit growth numbers for so many years now if they only ever cost money, realise no benefits, are just generally useless money thieving idiots? If you are a fly-by-night you can do it for a while, but SAP has been around way to long for that to still work. If only every business exec could read this post and realise the errors of their ways!
    I think SAP makes the best businesses run better.
    And remember: there is never only one side to any story.


  14. Hello. I am one the SAP Basis/Technical Consultant since 1994. My 2 cents advice is for Kiwi to hire its own Technical/Functional Internal Consultant who champion for the Kiwi Bank interest. Dont totally rely on external parties.
    Lesson learn from projects ….reallly paid off.


  15. Lance,
    it looks like you are a selfmade SAP and corebanking expert based on Gartner, Forrester, Celent, blablabla….

    I do corebanking and analytical banking with SAP for now 18years
    and I share every comment Protazy Mickiewicz says. Specially when he says “no more questions” after you stated that you have never been in a SAP Banking implementation. So you have got no experience on the ground just following what some competitors of SAP say.
    SAP is a specialised vendor of a corebanking and analytical banking software solution.
    The first implementationo was starting from 1995 in German DEPFA bank by CSC.
    Since then SAP CoreBanking modules are live in hundreds of banks, financial service institutions and corporate tresury all over the world (CML,BCA, TR-TM, SEM-PA, Bank Analyzer, SAP DM, Loans on Platform, CMS, Fraud,….)
    You should actually check the failing rate of implemetations of Temenos, Oracle FlexCube and all the other back dated corebanking applications.


    1. Large project implementations have multiple perspectives – from senior management, to the team of external and internal people on the project, and in my mind the most important, the end users.
      From what you and Protazy write you are in the second group – the team of people putting the system in place. That’s fine, but my own experience is that large project teams necessarily and often end up with their own internal culture and also drift away from the end user needs. They may downplay or simply not see their own often destructive impact on the organisation coalface during and after implementation. And I’m not getting too literary here – I’ve seen poorly implemented systems have a temporary or permanent devastating impact on the uptime of mining vehicles, that are in turn the bottleneck for extracting tonnes of ore worth tens of millions. I guarantee that this impact never makes it into the SAP implementation reports, while on the ground maintenance engineers scrabble to work their way around the system.
      Looking at the comments on this post and others, it’s clear that the end-users are often not happy, whether in SAP banking, other SAP modules, or, to be fair, in essentially any large IT project. The disconnect between end users and people inside the project teams is evident in the disconnect between the comments here between insiders and the end users.


      1. “That’s fine, but my own experience is that large project teams necessarily and often end up with their own internal culture and also drift away from the end user needs.”

        And whose fault is that? It is the customer who buys the product and it is he who should care about to learn it well. From my experience, people in organisations do not want to learn new things they are rather conservative and here it is, the answer to your question.

        I would like to underline on fact. You went from a “defective product” but ended up with a completely different opinion as I could state from your last comment. Which one is it then? The people or the product?


  16. Lance, End-users never like change, this is common.

    I agree with all of Protazy Mickiewicz comments. Others input, such as I saw a SAP implementation or I’m a retail payments expert and these are my views with no solid background or knowledge or expertise are baffling. Also calling it slow and not real-time. SAP has one of the fastest real time data products for such large scale IT systems – SAP HANA. (SAP HANA converges database and application platform capabilities, in-memory to transform transactions, analytics, text analysis, predictive and spatial processing so businesses can operate in real-time.)

    But with so many non-experts giving there opinion it was hard not to comment and make the majority realize to perhaps only comment or create debate on subject matters that you actually know about!!!

    P.S – No SAP consultant earns 5000 nzd as daily rate. I would know I I supply them to these project globally.



  17. On th internet, everyone’s an expert !
    (…but at least Lance admits that he’s not… before acting like one…)


Comments are closed.