There's no better way to derail a product organization than to accept to do a 'special'. A 'special' is a customer request that is not aligned with product roadmap and is done only because of the substantial financial incentive involved. The financial incentive is rarely huge enough to offset the total costs of delivering the 'special', but unfortunately the financial incentive is tangible while costs are mostly hidden in the lost opportunities and defocus of an organization.
At USV product summit last week I've heard yet another horror story involving a 'special'. A very well known USV porfolio company got an almost 7 digits big offer to do a special product for a top American brand, but one completely misaligned with the product roadmap. The business person responsible for the deal was blindsided by amount of dollars involved, so he managed to get the deal accepted by the leadership of the company (which was equally blindsided by all the zeros in the proposal). Once the deal got to execution phase it sucked all the time of half of the engineering team for three months. And since this become the most important project of the company, it sucked all the time of CTO, too. For three months, the company was focused only on fulfilling the 'special' and completely disregarding developing their real product. In retrospect, the company evaluated the costs of the 'special' to be substantially higher than revenue accrued, so they decided not to agree to any 'special' ever again.
I've seen dozens of such examples also myself. Marty Cagan has dedicated a whole chapter of his book to this issue and also wrote a blog post on the subject. So do yourself a favor and just say no whenever confronted with temptation of doing a 'special'.
This is part of series of posts on managing sales.