After your API is active, multiple calls from multiple users occur in short periods of time. How to define the outreach of your applications? What success looks like? To help answer these questions, we have defined KPIs (Key Performance Indicator) for APIs as indicators that can help us to measure progress towards your goals.
This technique is widely used by executives and bosses when managing businesses, enabling communication about the development of the company to other employees. Through KPIs, all employees are known and become involved in the mission of the corporation, in order to align the efforts around the strategies established. Therefore, the results pointed out in the KPIs quantify the company's performance and allow workers to understand how much their activities contribute to the success of these numbers. In short, KPIs are the measures that are used in the evaluation of a process.
Through Sensedia’s platform, a range of different information is sent to the product owner: number of calls, active applications, number of developers and thousands of other variations in the use of its APIs. What do you want to achieve with the API? Your answer can vary from “New business line” to “Accelerating internal projects”, we must think about how to define success in a product with so much data in almost all situations.And how to define success with these metrics? Initially you should ask yourself:
When answering these questions, you will notice that the API is affected by several areas, and you realize that important metrics for each of these areas can be relevant for analysis, as in the example below:
Looking at this table, I can say: there are multiple ways to measure what success looks like for your API. The origin of the calls may not be relevant to the channel metrics, or the uptime can be a very important factor for both the Service and the Structure. But how do you decide what matters to your API with so many options?
I will help you!! Here you can find 6 steps that help to define the meters of the current APIs.
To illustrate better the importance of the topic, I have an example of the following case: “After completing a successful expansion strategy, the executives of a fashion retailer wanted to improve the way they measure the chain's performance. Several stores had been opened in the recent past, in different markets, which made the operation more robust. To find out if a store was good or bad, they analyzed, mainly, the sales volume. If the store sold a lot, it was considered successful, and vice versa.”
However, looking at the image above, we find that this type of analysis is not enough, since, although a factor is significantly important (such as profit), it is modeled by several other factors that must be taken into account. Thus, it is important that you choose your lens, defining the priority points and which view will be used when defining the metrics and their type.
Get Know the Devs Funnel
The concept of a developer funnel is very similar to the concept of a sales funnel, and the sales funnel is a theoretical model that represents the path your customers go from the first contact with your brand to making a purchase. Its objective is to map the steps and define which ones can be used in each phase, to allow the advances of possible consumers towards the final goal.
So, if in the sales funnel the buyer's path should be mapped from the first contact to a purchase, in the developer funnel we should closely monitor how the developer interacts with the product, and for that, we will tell important values in each of the steps shown above:
Developer portals with concise and well-structured information encourage the developer to code something with their content.
Through these indicators, you will know if your API has a complete process with the developers who use it, bringing in new developers who will also go through your funnel.
It is important to check the treatment of your company's workflow. For example, you should look at the total number of devs consuming their APIs, which Devs are working on certain applications, which applications have specific groups of users, and how you analyze and spread that data. This information all goes through an ecosystem, like the drawing above. The ecosystem is very useful to think about the value of the analyzed data, to observe, for example, how devs react to certain changes in APIs, how users react to apps linked to these changes, and how you, who are looking at all of this situation, interpret all these values!
In the mid-1900s, Joseph Moses heard from the farmer and consultant Wilfredo Pareto a complaint regarding the proportion of beans generated by their pods. Thus, the Pareto rule was born, also known as the 80/20 principle, which says that approximately 80% of the effects come from 20% of the causes. Thinking about APIs, this means that few applications bring the vast majority of calls. Think about how it affects our API, and then use KPIs to boost who are the devs and apps with the most calls and how you're going to help them produce more.
How is the use of your API? Who makes the calls? Does the API have registered, more internal or external devs?
If you use your API don't forget the impact on KPIs, always take into account that internal values can change metrics such as income/dev.
Escape the Traps: Traps? To answer this question, I have a very simple example below:
In this example, it is visible how to be careful with the vanities in the data. It is important that there is a process for the metrics to be defined, and a system to analyze them, in addition to a feedback system so that the evolution is constant. Otherwise, you may end up making 6 times more calls to do the same thing, having unnecessary expenses without reaching anything more.
Finally, it is clear that “Who” and “What” should guide your KPIs, the importance of having a metrics definition process was also seen, and that if the lenses are not used correctly, you will end up thinking that only profits matter. For the measurements to evolve constantly and become more accurate, it is enough that KPIs are prioritized and the process of defining KPIs is periodically redone.
If you have any questions, contact us!