Hire the best consultant
It should go without saying, but in in consulting it applies to the extreme. A study about software developer productivity (Sackman, Erikson, and Grant (1968)) found that a good software developer can easily be 10x more productive than a not so good professional with the same amount of experience. There has been some, generally a minority of criticism, but the finding was in general well received, and explaining the “10x developer”, as it is known, is still a hot topic in the development community. While this study was done on software development, which is only part of consulting, it is still very relevant to consulting.
One of the explanations is not the amount of code produced, but the choice of what to code, and where to apply focus. If that is true, 10x applies even more to BI consulting.
There is an important thing keep in mind in consulting, whether it be BI of software development. The overwhelming benefit of a better consultant is passed to the client, not the consulting house. That is because consulting house is rarely able to charge even 1.5 times more for a good consultant with similar experience, than a mediocre consultant. The house at most benefits from indirectly from the effects of a happy customer. But the [n]x benefit goes to the site where the consultant is stationed. That means that good consultants come heavily discounted.
Hiring an exceptional consultant, is the best thing you can do to get the most from a consultant. If 10x is even remotely true, it far outweighs small differences in rates, convenient availability, and (with a few caveats) the tools used.
I had to be Captain Obvious. The truth is this does not get the consideration it deserves. It is more important (even from a cost perspective) than the right price. It is probably more important than the exact schedule.
Make your site attractive to the best
I’m not talking good coffee machines (that helps).
A good consultant will want space to innovate, and will favour interesting and challenging work.
Requirements: Be clear, but don’t over specify
The value objectives should be clear. A BI solution is not an automation of reports or saving a daily report creator time, but to provide analytics for better decisions. These values should be communicated, terse and clear. These are often one or two liners, such as “We are trying to minimise storage costs. We measure things like total storage cost, and cost as percentage of GP, and cost per day”.
Rather than, “this is our daily store costs report, that gets send to these people. We want to automate that, so we can save one person 30 minutes each day. We need these 20 columns in there”.
You pay big money for a consultant. Give him or her a brief and suggest something. A tool such as QlikView or Qlik Sense allows very quick prototyping, that can be modified if it does not hit the mark.
Get the right tool
You can have endless checklists, but the proof is in the pudding. See my blog on why Qlik is the best.
Don’t pay for consulting services to
- beg for data
- beg for server access
We do that from time to time. Having these sorted out goes a long way towards ensuring best use of time.
Don’t neglect staff empowerment
The solution is only as good as the hands it ends up in.
Staff empowerment is not looking over the consultant’s shoulder and understanding every bit of code. Qlik code is not compiled, it is sequential, not object orientated. It is transparent and it is not hard to understand.
Rather have your staff understand the philosophy of visualisation choices the consultant made.
Test drive the solution while it takes shape.
IT infrastructure staff needs to know what is required to keep the solution running.