Friday, January 14, 2005
The instant the legal types get involved with any consulting agreement there are huge chasms opened that immediately put the parties into an adversarial relationship. One of the classic tactics taken by the client is to demand liquidated damages be written into the contract, not hugely unreasonable in most cases, depending on the exposure, but in other cases, especially with large financial systems that interface with the marketplace, the potential exposure can be astronomical for the consultancy, and as such, they will either defer entering into a final agreement and start the work anyway, or implement a risk management regimen so restrictive and so argumentative that it will place a huge burden on the engagement team before any real work is done. This problem is usually compounded exponentially by the perpetual conflict between the client's desire for a fixed price engagement, and the consultancy's insistence on T&M. Usually, there is some compromise in the form of professional day billing, but being brutally honest about it, most large consultancies are notorious for underestimating the effort involved in a project (at least as far as what the client hears; I've been party to any number of internal discussions where for example, we would say it would take 24 man-months to do something, and the sales guys were insisting that we bring it down to 18 man-months to meet the client's price point). This inherently leads to "Work Late, Bill Eight", and given the unrealistic productivity targets assigned to many consultants, leads to a disaffected team that is liable to defections at any moment.
The problem of course is hugely magnified when a project is part of an overall program, and centralized decisions and standards are being applied to something that may just be a one-off. It's bloody well impossible to apply Six Sigma standards to something that is not productized, although perhaps individual components may be. While over the long term, statistics and metrics prove valuable and might even provide some targets to achieve, sometimes there's just no substitute for someone sitting there and manually verifying all of the work that's done. The trader who picks up his auto-ringdown damn well better get the broker-dealer on the other end, and he won't give a crap if someone screwed up a cross-connect or whether it wasn't verified by the person checking the turret when it was programmed. He's the profit center, and he's going to scream bloody murder. ISO 900x be damned, it better be right the first time. Squawking about metrics isn't going to replace good old fashioned due diligence. Unfortunately that means checking the checklist over and over again until just before go-live, and that's an expensive proposition, one that Sanjay and Apu can't do remotely from Bangalore. But it still doesn't mean that the consultants are necessarily going to get paid in full or on time for all of the over and above stuff that gets done. Clients know they're in the driver's seat, and unless you're talking C-level dealings (and even in that case that won't necessarily help, for example, the CEO of Colditz cordially hates Respected Employer as an entity) there won't be any easing of consultant-client tensions. And unfortunately due to shareholder pressures, that means contracts will be written to squeeze consultants, and Accounts Payable will have a labyrinthine maze of approvals designed to put anyone's 30/60/90/+ receivables schedule into the toilet.
And the hilarious thing is that consulting firms end up using their consultants as credit departments. I well remember being pulled into an AR situation, where I'd sold a very minor gig to Colditz (this was years ago, long before The Program From Hell) with only one consultant (not moi being staffed). Unfortunately, a third party was involved, as the specific project was really nothing more than a staff augmentation for something the third party was doing. I sent the young man downtown, and he did his work happily and productively, and I'd heard no more about the matter for months. All of a sudden I get a nastygram from AR where the invoices hadn't been paid for the project, and would I please collect from them. Not being Dun & Bradstreet (and believe me, Colditz and the third party are huge, NYSE-listed firms in very good financial condition, with huge bureaucracies) I waded into trying to resolve the situation, where I had two huge firms pointing fingers at each other saying "Invoice them!" (meaning the other firm, of course). It took several months to resolve, but after a while the check for the young man's first month's work came in. Twelve months after he started. It was several months before they were current (it ended up that the third party paid, rather than Colditz). Despite pointing out that I'm involved in client service, the AR department utterly didn't care and didn't want to do its own due diligence to track down its money.
On the more immediate front, we tracked down an SME in the technology we're looking at for Esteemed Client, and he said, well, yes, on paper it'll work. It's a complete contravention of the vendor's ideas and certified means of using said technologies, but it'll work. On paper. And that assumes that Esteemed Client can change its business processes (fat chance!) to make sure that certain long-running events don't get kicked off at inconvenient times. Sales Guy might just be thrilled. I still don't think it'll work from a practical perspective, and on top of it all, I have to play some games with load balancers and DNS that I really didn't want to get into, at least at this point. Sigh.....