An Introduction to Setting KPIs for Your API Strategy Posted in Strategy Thomas Bush February 4, 2020 If you work at an established enterprise, chances are you’ve already heard of Key Performance Indicators or KPIs. These objective metrics offer stakeholders an effective way to measure the success of business initiatives, and now they’re working their way into the world of APIs. In this introduction to KPIs for API strategy, we’ll discuss how and why you should set KPIs, and how you can get the most out of them. This post was inspired by a talk given by Richard Jones of Dun & Bradstreet at the 2019 Platform Summit in Stockholm, Sweden. Although KPIs were just a small part of the presentation, we highly recommend you watch the full video for a range of powerful insights on practical API usage in the modern enterprise. What Are KPIs? Key Performance Indicators (KPIs) are tangible metrics that reflect the success of a business initiative. They’re a common sight in the business world, giving professionals a simple, objective way to evaluate performance. It’s worth noting that KPIs can be both company-wide or operation/department-specific. As APIs become a more significant part of modern business strategy, it only makes sense that they, too, would be subject to their own set of performance indicators. Popular KPIs for API platforms include API revenue, number of APIs, Time to “Hello World,” and the number of active developers — but we’ll talk about which of these you should and shouldn’t care about later. Why Set KPIs? In his talk at the 2019 Platform Summit, Richard introduced the primary benefit of KPIs with the age-old adage what gets measured gets managed. It doesn’t take much first-hand operational experience to know that this is absolutely true: if you’re continually monitoring a statistic, you’re a lot more likely to take steps to ensure it moves in the “right” direction. For example, if you regularly measure how many people are calling your APIs, you’ll probably take steps to increase that statistic. A secondary benefit of setting KPIs is that they provide a consistent frame of reference for the performance of your APIs, accessible to both business and technical folks alike. As we’ll discuss in just a second, well-selected KPIs should closely reflect your business strategy; ultimately, this gives stakeholders an objective way to see how APIs are performing in areas meaningful to them. Related: How to Measure the Success of Developer Relations Choosing the Right KPIs There are dozens of KPIs to choose from thanks to the digital, data-laden nature of the API industry. Number of APIs, the number of customers, speed to API, traffic, revenue, and cost reduction are all popular choices, but how do you decide which KPIs you should really be looking at? The answer is simple. The primary KPIs you set for your API strategy must align with how APIs add value to your business. As a result, you ought to avoid so-called vanity metrics (such as the number of APIs or number of API calls), which might sound impressive, but don’t always correlate to business value. Instead, go back to square one and ask yourself, “how are APIs beneficial to our business strategy?”: If your APIs are built to drive revenue, API revenue should be a KPI. If your APIs are built to grow awareness about your organization, the number of developers, customers, and/or partners should be KPIs. If your APIs are built to reduce costs, money saved by not using third-party APIs or needlessly recreating functionality should be KPIs. Of course, some KPIs are easier to track than others. For example, measuring developer sign-ups could be as simple as tracking the number of API keys distributed each week. On the other hand, estimating how much money an API has saved your organization might require the use of complex financial and operational models. You might be wondering: what about all the performance-related “KPIs”, like API call latency or server error rate? Are they really KPIs? Unfortunately, this seems to be a matter of semantics more so than anything else. For some, performance figures encapsulate the very literal meaning of a Key Performance Indicator; for most, however, the term KPI is strictly used to refer to business metrics. In other words, performance figures are important — and should definitely be tracked — but they aren’t necessarily KPIs. Also read: Using API Analytics to Empower the Platform Using Your KPIs Getting the most out of KPIs takes more than just setting them — you actually need to use them! The easiest way to do so is by putting KPI figures in front of yourself (and stakeholders) on a daily, weekly, or monthly basis, and let the principle of what gets measured gets managed to do its work. For this reason, Richard suggests you auto-generate KPIs. This starts with choosing the right metrics: instead of focusing on qualitative data — which usually requires manual interpretation — leverage the mass of quantitative data already available to you and have code do the heavy lifting. Aside from just reviewing KPIs on a regular basis, you may also wish to set goals for them. For example, if the number of developer sign-ups is your primary KPI — and you usually see 200–250 developer sign-ups per week — you might set a goal to increase that average by 20% within a six-month timeframe. Finally, use KPIs as dependent variables whenever you test the effectiveness of an individual strategic decision. For example, if you’re deprecating functionality from a core API, monitor how this change affects your primary KPIs (e.g. API revenue) to know whether you’ve made the right decision in the context of your greater API strategy. Also read: 7 Steps for Building Successful API Products Final Thoughts KPIs offer a straightforward way to track the performance of your API programs. While there are hundreds of metrics out there that might masquerade as KPIs, the truth is that good KPIs should closely reflect the business value APIs provide to your organization. This means that API revenue, savings, and sign-ups are the truest forms of KPI. To get the most out of these metrics, make them easily available and use them as your frame of reference for goals and experiments.