DataFlux - The Leader in Data Quality and Data Integration

What Consultants Don’t Tell Clients

Phil Simon

May 6, 2010

In my last post, I discussed what clients don’t tell consultants on data migration projects. This time, I’m reversing things. This post will focus on the major things that consultants don’t tell clients.

Disclaimers

Before continuing, a few disclaimers are in order:

  • While not my sole source of income, at least two-thirds of my annual income comes from consulting and this has been the case for nearly ten years. Translation: I tend to write from the consultant’s perspective. Do with that what you will.
  • I’m going to include the salespeople here from system integrators (SIs) and consultancies. While they are not consultants per se, most clients justifiably do not draw the distinction between these two groups when data migration and integration projects run amok.
  • I don’t believe that I am inherently biased against salespeople, a topic that I recently addressed in an interview. I just have seen certain shenanigans more than once before on the sales’ side.
  • I’m obviously writing here on general terms and from my own–hopefully not too jaundiced–eye. If you’re a consultant or salesperson, you may be reading this and thinking, “I would never nor have never done that.” I believe you.

Unaddressed Questions

In any economy (much less a challenging one), SIs have an incentive to overlook potentially showstopping  data and people issues when, quite frankly, they ought to know better. This may take one of several forms:

  • Failing to ask key questions during the sales cycle. These include:
    • How old is the data?
    • Has the data ever been profiled?
    • How many sources of information need to be cleansed and ultimately loaded into the new system or application?
  • Opting to overlook answers to these very questions that will probably inflate the cost of the engagement (and risk not getting the deal).
  • Inappropriately tweaking estimates to lower immediate costs, knowing full well that the SI will submit change requests later in the project once the client’s checks have cleared.

Now, when a project goes awry and things get contentious, it’s simply wrong to instinctively blame the consultants and salespeople. While circumstances vary, in many cases, there’s simply no way for the SI to have known the severity of the data and people challenges at any particular client. No project estimating tool can accurately account for the cauldron of  issues endemic at many dysfunctional organizations. It’s unreasonable to expect a SI to send in a team of consultants to do weeks worth of data profiling and chalk the costs up as “pre-sales.”

Simon Says: Advice for Clients Evaluating Consultancies

For clients interviewing prospective implementation partners, understand that even estimates with the best intentions are just that: estimates. Depending on the way that the contract is ultimately structured (T&E vs. fixed bid), expect your new partners to behave accordingly. If the client opts for the former, then all issues will probably be addressed and hopefully resolved–on the client’s dime. On fixed bid projects, consultants are incentivized to complete the work as quickly as possible and move to the next client. Understand this going in and make decisions about error resolution accordingly. There is such a thing as tomorrow, especially regarding nonessential data.

Be realistic and transparent. CIOs and their teams should not expect consultants to be clairvoyant with respect to data issues. Nor are consultants miracle workers. Clients should know that SIs have their own predispositions–just like they themselves do. Don’t hide key problems for the purposes of securing the best overall project bid. You’re looking for a partner, not a scapegoat or punching bag.

tags:  , , ,

  1. #1 by James Standen at May 6th, 2010

    Great post Phil.

    On the data profiling front, while I agree with you that its unreasonable to expect a comprehensive review, I think it is important that some analysis is done.

    Even just a day with a good data profiling tool and the data will give insights that will most likely improve the accuracy of any estimate significantly.

    But the best approach for large projects is probably to do the project in stages, the first stage being data profiling- and have the result of that study available to bidders, so when the main project is estimated it is done on a consistent basis, and with some knowledge of the underlying issues.

    Of course, even after data profiling, count on finding more things as the project unfolds.

  2. #2 by Phil Simon at May 6th, 2010

    James

    Completely agreed. Unfortunately, at least in my experience, no one ever does any data profiling at any stage. While this would make a great deal of sense at the beginning, often things are already stretched budget-wise. People often consider data profiling a superfluous expense when, I’d argue, it’s anything but.

  1. No trackbacks yet.

Comments are closed.