Preventing Premature Product Death

For many of us, our products are like babies - we nurture them into being and we carefully tend their growth. And while we engineer them into being, they are still highly susceptible to never reaching their potential. We try to craft tight bundles of elegant utility whether we're building devices, software, services or even high-tech clothing. It's scary. It's risky. It's costly.

With TVs, phones, enterprise reporting, medical devices or wearables, there's a tightrope we walk. A few missteps and we lose control of that line where we open ourselves up to Feature Creep. And have no doubt, feature creep is among the leading causes of product death across the globe. Even the best and and brightest can get caught up in its clutches. Don't let your product be the one requiring a *fire* sale.

Des Traynor of intercom.io offers us this matrix on making the crucial feature decisions. It's important to know how many people a feature serves (x-axis) and how important it is to them/how often they use it (y-axis).

“One way to think about this is this (he shows the slide): the core of your product is here (points to top right), you make an improvement here everyone appreciates it and it pays off big time. You’re in Dangerville over here (points to top left), right? Small number of people heavily dependent on small PC or product. The reason I call this Dangerville is because it’s the road to consulting-ware” (as in software made by consultants for clients). - Des Traynor, intercom.io

This logically leads you to creating a more effective way to rank features. If each box is ranked 1-4 and 1-4 and then multiply them, you'll get a natural prioritization - where your 16's are must-haves, your 12's are good to have, your 6's should be second or even 3rd release, if at all. Your 3-4's may concentrate in the few places where you can command a high premium for a specific feature bundle (such as camera photo enthusiasts or high tech mountain climbers) or you can consider launching a highly specialised product, provided the segment is large enough (at this point often in the 100,000's of units, not millions).

This approach also allows you to align your organization's talent effectively and determine how many and who should work on what. An individual whose talent lies in a feature set (i.e. augmented reality) that might not have a broadscale applicability (yet) but might be extremely valuable (in this case potentially to segments of architects and geologists). He can be a high value individual contributor, but not likely equipped with a full team and budget largesse* who manages market analysis, licensing and vendor evaluation. That probably also means he needs a different skill set than our traditional product engineer.

As organizations branch out into ecosystems - where designers, engineers and even product managers are hired guns, this feature score matrix provides a valuable approach to helping management put the right team in place to deliver new products that have enough greatness to get up on their own and walk boldly into the marketplace.

The idea that started this for me is here: thanks Jason.

PS - a lot of matrices will use 5 boxes where three is the middle of the road on each axis. Resist that temptation. It makes an easy place to decamp a bunch features as 9's (3x3). I'd recommend strongly against that and put in a forcing function. If you don't begin to create a firm line, you won't get very far.

* I know, big budgets are the moral equivalent of unicorns where we all want to believe they exist. I have this vague recollection of them in the past, but it might all be fairy tales by now.