Unless you have a direct line from feature to revenue the impact of a given investment can be mixed in with hundreds of other factors that are hard to differentiate.
You can probably measure how much is your org spends on R&D, engineering and product management. If you’re reading this, I’m going to guess it's in the millions or tens of millions per year. So that's your ROI denominator.
Attributing returns, though, can be tricky. Did a feature result in a sale? How about if it prevented potential churn – how would you measure that?
Establishing KPIs for every launch
One best practice is to establish written, and agreed-upon, KPIs for each launch of a certain size. The KPIs should be in the form:
- By X days after launch
- Y should increase/decrease to Z
- Based on the measurement in system ABC
An example: By 30 days after launch our average customer acquisition cost from paid search should decrease by 2% based on Google Analytics.
Presumably, KPIs can be linked to business results, and from there a view of ROI.
Establishing a North Star metric
For some businesses, especially those with large volumes of customer interactions, there may be one or more metrics that sum-up the overall health of the entire platform. I spoke to a CMO of a major jobs website who told me that everything in their business is based around a single metric they called "employer engagement". She then measures the ROI of every launch based on the delta between before-and-after launch results against that single metric.
Other "North Star" metrics might be revenue per customer, return on CAC, or NPS.
Improving the ROI of Launches
While measuring the return on product development isn't an exact science, directionally there are clear ways to improve it.
Early in my career I was the manager for a product I inherited that was severely challenged in the market. It was under-powered compared to established competition and it felt like every client meeting brought up another set of missing capabilities that would prevent adoption.
My engineering team started rapidly knocking off the most important gaps and we saw some momentum. But then I heard the bad news, a major prospect wasn’t going to use our product this year. Why, I asked? Well we can’t bill them on a flat rate, the sales team told me. But, I had built that exact feature last month! A sale was lost for no reason, and the product investment gone to waste.
I can almost guarantee you that this happens in your organization at some level. A customer churns because they’re not getting the most out of the product. A prospect puts off their trial because they don’t perceive your solution the right way. A support ticket isn’t answered properly, resulting in an NPS decline.
So how do you make sure this doesn’t happen? There’s no one answer, but from our research with hundreds of product marketing organizations, we found some commonalities.
- Coordinate all your product changes, large and small, with the commercialization process. (That’s why in LaunchScience we built Jira and ProductBoard integration)
- Use standard templates for different launch sizes
- Communicate often and actively
And even if it isn’t entirely scientific, measure what you can and try to get a view of ROI.