Wed 14 Dec 2005
The sad state of Analytics Consulting
Posted by Nishith under BI, Data Mining, Analytics[3] Comments
Probably because they are creaming the market anyway, consultants from the bigshot analytics and BI companies are a royal pain to work with.
What we want from a million-dollar implementation project is the end result that delivers tangible value. And as that smilling pre-sales guy convinces us to sign that million-dollar contract, this is what we are thinking with a nagging gut feel that something might be amiss:
And this is what they turn out to be!!!:
Without taking any names, early in my career I had the misfortune to work with some who brought in a lot of hype and attitude without any deliveries.
As a client, I want the consultant to train me (if needed) and help me get up to speed. And I do not find it very funny if the project is in troubled waters.
In my first project with the industry leader in BI software and its ‘consultants’, I was naive enough to be surprised when I found that these guys were just trained on the tool features and had no perspective or knowledge on the what and why and how of Analytics. On cutting through the daily dose of hype that I was expected to be satisfied on, I found that the consultants were using a data value from the future to predict the same future!! If you are trying to predict for Feb 2006 in Dec 2005, you cannot depend upon information that would be available only in Feb 2006!!! It doesnt take a rocket scientist to figure that one out!
If only they had let me into their arcane art instead of putting up smoke screens, maybe I could have spotted this goof up a couple of months earlier.
Maybe some of these guys would read what it takes to Create Passionate Users:
Creating Passionate Users: Passion is blind
(Note: The images above are from Creating Pasionate Users Blog by Kathy Sierra, the co-creator of the bestselling Head First series.
January 28th, 2006 at 9:58 am
Oh, those f***ing vendors! The most lucrative part of the IT Con.
Did you force them to come in and do a proof-of-concept on your systems? To be fair, keep it limited in scope, but never let them drive it.
I’m still amazed at how unwilling companies are to bring in independent consultants to analyze their needs objectively and work with them to find a vendor. Is it really a bad investment to spend $25-$50k for help before signing that million dollar check?
Vendors will never, never tell you the WHOLE truth, even if they understand it, which they don’t.
I just went thru a vendor evaluation with three global vendors and each of them sent in idiots that they apparently thought would be good enough. It took less than a day before the sales reps were flying in more “experts” from around the country.
Vendors can go thru a lot of turnover because their consultants can make more money working for companies providing support than working for the vendor themselves. Try to find a consultant from the vendor that has 5 years or more experience WITH that vendor (not just with some of their products).
Vendors suck up the costs of POC’s (unless you buy, then they suck it out of you in the deal) and it onlymakes sense that they try to do that as cheaply as possible.
I finally found a product I liked, but I still had to put up with their consultants and still got screwed a little with hidden costs, but that’s what they do. Just make sure you find out the list price and never pay more than 1/2 that (except for small purchases).
From what you say, I bet I have a good idea which industry leading BI company you had to deal with. If so, the software is a lot more functional than the boobs they send out know about.
January 29th, 2006 at 1:33 am
*LOL* Dratz, I wish you were blogging then. I should have raised certain apprehensions during project kickoff itself.
Unfortunately I had to step in as the Project Manager just 3 days prior to the kickoff. The deal had been inked and the project duration had also been fixed without even a requirements document in place. So it was crisis management for me from day zero. The good part was that this also meant I had a lot of creative freedom on the project. High risk, but fun.
In general I think its probably better to get the tool from one vendor, and then get another vendor for the implementation. Consultants from the tool vendor are almost always well trained on the “menu-options” but quite clueless on what the menu-option really does and when it should be clicked.
I agree that the “industry leading BI” software is quite functional (my one line project update to our CEO once read “Good tool, bad team”), but it does manage to piss me off sometimes. Blatantly peddling a script as a “programming language”, they should have probably revised their architecture a decade back (that’s easier said than done, you don’t want to risk too many changes with a big base of users used to the way things are).
Having said that, over the past few years I have been looking very closely at some of the very impressive work being done in open-source. In particular you should look at R project (http://r-project.org) which is based on an original design by AT&T Bell Labs and is amazing. I intend to write a lot about it in the coming months.
Having seen it happen, I am quite particular about ensuring that our consultants should not turn out the same. Analytics software market is getting commoditized, and consistent quality will provide the only sustainable competitive advantage for the analytics vendors.
January 29th, 2006 at 6:09 am
Amen to that. Commoditize all of IT and not only will quality and prices stabilze, it will be easier to find true value.
I’ve been working on my own opensource BI, but it may be smarter to join into existing projects. I’ll check them out.