peopleware-blogShallow2.jpg

Coaching

We’ve just posted an update on training and have an example of how we use some software to make easy to follow video user guides. You can find the new page here – or watch the sample system[read more]

For the last few months, I’ve been working with a brilliant coffee vending business near Gatwick. In terms of culture and atmosphere, it’s quite a contrast to BOTH other venture capital[read more]

The results of Google’s “Project Oxygen” have been reported all over the place in the last week. The London Evening Standard and the New York Times both carried this, follow the[read more]


The cost of complexity – part 2

I posted a while ago about the cost of complexity, and two events in the last week had me looking back at what I wrote then. The first is a client who DID end up adding a second remote access solution and the other who is trying to decide what to do about Business Intelligence (BI) in the face of a merger.

In the first case, the client agreed to implement an extra remote access solution in response to genuine staff need and its evolving business model. A small pilot was followed with a wider roll-out. The very narrow need of the first users was fine – a specific solution to meet a specific need. But in a short time, this became a more widely demanded service, and the additional users had not bought into the very specific operating model of the additional service. Complexity, problems and costs started to go and user satisfaction went down.

In the second case a client’s previous CFO had worked with a fairly niche BI supplier to build some financial and budgeting BI tools with a limited scope for Self Service BI. The problems were that the Extract Transform and Load (ETL) processes were very fragile and were not supported by the small IT Function. As data changed, the ETL and BI systems broke or became associated with doubtful data. A separate Sales BI platform had been implemented but was under-licensed for it to be used company wide. What is worse is that the two sets of headline sales data (from Sales and Finance) often disagreed.  So two BI platforms, neither ideal but due to the ongoing merger, neither is likely to be a long term solution. In the face of the need for different data from the new parent company, the temptation is to find a third BI Tool that is easier to manage to deliver the newly required information.

Using the table from last time, we can see the cost effect of the extra solutions – even if we assume the total cost of each BI Platform is as little as £50,000.

Incremental cost No of solutions Multiplier Marginal Cost of this solution Average cost per solution Total cost of all solutions %age increase over the single solution of this solution 5 year cost of adding this solution
£50,000 1 1 £50,000 £50,000 £50,000 100% £250,000
£50,000 2 1.33 £83,000 £66,500 £133,000 166% £415,000
£50,000 3 1.66 £116,000 £83,000 £249,000 398% £580,000

So, even though sticking with the less than ideal solution, the true costs of the 2nd and 3rd show that it will almost certainly be worth completing the work required to make the ETL (Data load process) robust, to use the for ALL the financial and sales data and deliver a single supportable solution that should serve the client well for the next 2-3 years. Criticially, at a time of change it also means that as the new parent’s requirements become clear, there is only one place to change the definition and use of the data – not 3!