Product Management (Part 2): “Do all features have ROI associated with them?”

In the last post, I took an example of a feature where we had a reference data. But invariably we run into features where it’s difficult to assess the impact when you do not have any reference or top line impact is not straightforward to determine. To give an example, let us say we want to introduce user data backup or having a projector built in tablet. One may say its common that these kinds of feature help the business, but truth is business decisions are not made on common sense they are driven by numbers (Will not touch the management gut feel in this post).

Let us look at the situations where we typically run into such scenarios. I would broadly classify features into following buckets,

• Sanity
• Insurance
• Opportunity
• New to World
• Marquee Customer
• MP (Management gut feel)

Now let us look at the details with example and what are the simpler ways to do business justification in following cases. These are the base rules, which you can use to derive the numbers before you do the feature and later measure the relevant performance numbers but if you are looking for the precision of last digit then there are better ways

Sanity features: these are must have features for product to compete in the segment. E.g., MVP (Minimum Viable Product) features, features inspired by competition (If you don’t introduce them then it might erode your user base). The ROI measure here is to use cost or loss for not having the feature in time. It can be in terms of user base loss & associated revenue, delay in scaling up or Time to market.

Let us take an example of chat feature that was introduced in classified apps. Once this feature was introduced by one classified it invariably became a sanity feature to compete in the market. The measure of ROI in this case can be derived from loss in number of transactions or decelerating growth rate (app installs, listings, new user signups, visits), which is encountered in, say 1-2 weeks.  

Insurance features: you can live without them but you are bound to fail without them. Examples include security, data backup and high available infrastructure. The ROI measure here is a combination of the bottom line impact as well as the top line. The way to derive is to use the probability of failure or risk factor (as business scales, the risk factor increases) and user impact because of it. If you are in B2B you have SLA (Service Level Agreement) data to help derive numbers else use the expected loss numbers in terms of transactions/users/revenues etc.,

An example case for this is what happened last year with a major event run by one of the top e-ecommerce companies. During the event many encountered issues like system down, order not placed, payment not working, multiple orders etc., The ROI in such case could have been worked out based on risk factor (for each of the modules), Expected number of transactions, Number of transactions which would be lost, Order value associated with them vs. cost of putting up things in place. One more so-called intangible factor, which you can put up here, is the customer confidence impact.

Opportunity features (More of B2B but applicable to B2C as well): these are the features whose absence would result in loss of opportunity. It can be business opportunity or market opportunity. Examples are 508, FISMA or any of the compliance, support for particular database/software or even a payment gateway etc.,

Example cases for this are extending your software to other countries where currency localization is required or extending your software to Healthcare or Federal sector in US, which have unique compliance certification requirements. The ROI in such case can be calculated based on Total achievable market, target market share (something that we plan to address) and probability of acquiring the customer vs. cost to build the feature.

Example for B2C case can be integrating with a particular wallet provider. The ROI can be calculated based on Total achievable market (users who exclusively use the wallet), percentage of users who fall into intersection of your target segment and wallet users, probability of transactions by these users and the charges associated vs. the cost of adding it. Another example can be getting a particular service provider listed on your marketplace.

Will take up other 3 scenarios and ROI calculation for a distributed feature in next post.