Prioritize your features by customer need

Prioritize your features by customer need
Josh Smith — Founder at Sift

What features should you build next? Start by identifying your biggest opportunities.

At any given moment, you have an infinite number of directions you could steer your product. If you're like me, your list of feature ideas is long enough already. You're not short on ideas. You just want to know which ideas deserve your focus.

Make it measurable

Lord Kelvin, the physicist who discovered absolute zero, put it best:

I often say that when you can measure what you are speaking about, and express it in numbers, you know something about it.

Although you can't achieve the level of exactitude or certitude that Lord Kelvin could in physics, you absolutely can make opportunity measurable.

What do I mean by opportunity? Opportunity is found in unmet customer needs.

Opportunity graphed with high importance and low satisfaction.

Imagine the perfect opportunity.

The customer need is so critically important that, on a scale from 0 to 100, they top out at 100. But the customer is so completely unsatisfied with how their need is currently met that they bottom out in satisfaction at Kelvin's absolute zero. You couldn't even graph it above; the graph would be indistinguishable from the background, a mere point without area.

If this were a "feature" in your product, it would be failing absolutely. Or if you were evaluating competitors and discovered this unmet need among their customers, then congratulations: you just found a chink in their armor that you can turn into a competitive advantage for yourself. Go meet this need for their underserved customers.

Your inner optimist should see this for what it is: opportunity.

Let's make this practicable by separating out how to assess both importance and satisfaction, and then how to combine them to measure and then rank your opportunities.

Assessing importance

Hopefully you already have some sense of the array of outcomes that matter to your customers. If you don't, then you badly need to run some customer development interviews and gather a list of outcomes. Jobs-to-be-done interviews can help you come up with a list of your "jobs" that customers will hire your product to do for them. I recommend reading Intercom's primer on jobs-to-be-done.

Once you have a list of these jobs, you can ask your customers to assess the importance of each. I recommend that you ask something like:

How important is X to you?

And provide the following choices:

  • Very important
  • Somewhat important
  • Not important

You can give each of these a numeric score of 1.0, 0.5, and 0, respectively. The scale you use isn't particularly important. But using a scale from 0 to 1 will make it easier when you calculate your opportunity scores.

If you have jobs that consistently come up with a Not important rating, then it's not a real outcome. Remove it from your list and move on. And don't even bother asking for a satisfaction rating from someone who rates a given outcome as Not important. Why waste their time?

Assessing satisfaction

With your importance ratings in hand, you can then ask about satisfaction, similar to how you asked about importance, something like:

How satisfied are you with how this works today?

With the following choices:

  • Happy
  • Neutral
  • Unhappy

As you did with importance, you can score happiness on a scale from 0 (unhappy) to 1 (happy).

If you're designing a survey, it makes sense to follow up each importance question with the satisfaction question, like so:

  • Feature 1

    • How important is this to you?
    • How satisfied are you with how this works today?
  • Feature 2

    • How important is this to you?
    • How satisfied are you with how this works today?

Identifying levels of unhappiness is particularly helpful, not just because you'll identify unmet needs that you can eventually turn into opportunities in the long-term but because you can use this data to have customer success teams follow up with the customers in the short-term. Reach out to them and figure out how you can make them happier.

Simplify your scales

Keep your scales for possible answers relatively simple. There's no reason to complicate this too much.

This is especially true since you're likely to have a fairly long list of possible outcomes that you want to mine for opportunities. Think about just how much work even a simple product does. Even the most basic features in your product like authentication, account management, exports, notifications, and so on will, combined, contain a laundry list of unstated customer needs.

Since you'll have such a long list of outcomes, you'll need to ask your customers a pretty long list of questions. If you want a high completion rate so you can have more data to act on, then you'll need to lower the barriers by asking questions with fewer choices, as opposed to some 5-, 7-, or 😱 10-point scale.

Combine importance and satisfaction to find your opportunities

Ultimately we want to satisfy the needs that customers find most important.

You want your feature to be important and your user to be satisfied.

Brace yourself for a little math. We're going to break down how you can calculate total possible customer value, the customer value delivered, and the opportunity to add new value. Let's brush up our multiplication and subtraction skills.

Calculating customer value

Consider the following set of features:

  • Feature 1

    • Importance: 0.9
    • Satisfaction: 0.3
  • Feature 2

    • Importance: 0.2
    • Satisfaction: 0.8
  • Feature 3

    • Importance: 0.7
    • Satisfaction: 0.7

We can calculate the value delivered by multiplying importance and satisfaction, which yields the following formula:

Customer Value Delivered = Importance x Satisfaction

You could graph this by shading in the area (in blue, below) bounded by importance on the x-axis and satisfaction on the y-axis.

Customer Value Delivered = Importance x Satisfaction

This is a visual representation of the customer value delivered, expressed as a fraction of the possible value delivered by a hypothetical perfect feature with maximum importance and satisfaction.

We can derive the *customer value- delivered for all our features:

  • Feature 1 – 0.9 (Importance) x 0.3 (Satisfaction) = 0.27
  • Feature 2 – 0.2 (Importance) x 0.8 (Satisfaction) = 0.16
  • Feature 3 – 0.7 (Importance) x 0.7 (Satisfaction) = 0.49

Comparison of graphs

You might be tempted, if you focused solely on satisfaction, to think that Feature 2 delivers more value than it actually does. Your customers are really happy with the feature, right? But its low importance actually devalues the feature; even if they were wholly satisfied the value would only ever amount to a score of 0.2!

Calculating opportunity to add new value

In contrast to our low importance feature, let's look at Feature 1. We know the total possible customer value created by Feature 1 (0.9) is equal to its importance (also 0.9). You can't make an outcome more important to a user, but you can determine how satisfied they are. So an importance of 0.9 could be totally satisfied and deliver a total value score of 0.9.

Graph of the opportunity score

Our feature has a lot of unmet value, or opportunity to add value. The formula for calculating this opportunity is:

Opportunity = Importance x (1 – Satisfaction)

Plugging in our values for Feature 1, 0.9 x (1 – 0.3) = 0.63. You would get the same value with the equivalent formula:

Opportunity = Possible Customer Value – Customer Value Delivered

This makes sense when you think about it from the perspective of that total possible customer value: it's simply the combination of the opportunity and the value delivered.

Graph of total possible customer value

Applying this formula to all our features, we can calculate all our opportunity scores:

  • Feature 1 – 0.9 x (1 – 0.3) = 0.63
  • Feature 2 – 0.2 x (1 – 0.8) = 0.04
  • Feature 3 – 0.7 x (1 – 0.7) = 0.21

Comparison of graphs

The lightly shaded areas show the opportunity to add value for each feature.

We can already see that Feature 1 has a massive unmet need (or opportunity), especially in comparison to the other features in our list. The difference is even more visible if we overlay the shaded areas on top of each other.

Graph of opportunity scores compared

The opportunity score for Feature 1 is nearly 16 times that of Feature 2. Which one would you focus on?

But lest you think we're done, remember that not all features (or outcomes) should be treated equally. Importance and satisfaction simply don't exist in a vacuum where there are no other product considerations.

Consider your coefficients

You can't consider the opportunity score alone when deciding priorities.

Not every feature you want to build has the same complexity. Clearly not everything you do will require the same amount of time or the same level of effort.

Multiplying your opportunity scores with a coefficient like time, effort, confidence, or any other factor you find important will help you to weight your results appropriately. You can even use your opportunity score in something like the RICE prioritization framework introduced by Sean McBride at Intercom.

Rank your results

Once you have your weighted score, you can rank your results to determine your highest product priorities.

Of course, you may choose to do things out of order or choose to intuitively weight your priorities even if you can't reduce your reasoning to some numerical value. This framework is not intended to restrict you, but to be used as a guide.

With our three example features above, assuming they required the same amount of work to improve, I would first focus on Feature 1, then Feature 3, and only then work on Feature 2. In fact, Feature 2 appears to be so unimportant that I might even experiment with killing it.

Download the spreadsheet to calculate opportunity scores

You can use a spreadsheet to calculate your opportunity scores once you've collected and averaged your scores. Feel free to copy this spreadsheet for yourself.

Request an invite to Sift

Sift is invite-only right now. Sign up to be invited.

Icons made by Freepik and Smashicons from Flaticon are licensed by CC 3.0 BY