APIs are on track to become invaluable assets in a growing number of companies. And API metrics are more crucial than ever, as they help identify issues, optimize performance, and ensure API consumers that their applications are delivering the expected results. This blog will look at API metrics, on both a technical and a business perspective. Don’t treat them differently, they are all play a part when it comes to growing your API adoption. You will also see how e-commerce-like analytics can help you improve your API business.
Technical API Metrics
On the technical side, API metrics help developers assess the performance and behavior of APIs. These metrics can provide valuable insights into how APIs are being used, how they are performing, and how they can be optimized.
What is the difference between Response Time and Latency?
Before diving in, let’s take a look at those two terms, which are often used indifferently.
Latency measures the time taken by both the request and the response to travel between the API consumer and the server. These metrics mostly depend on the use of a proxy — which tends to have a different impact according to the type of platform — and geographical contingencies, whether the requested server is very far away or just next door. For example, Blobr latency is up to 100ms.
Response time is the time it takes for an API to respond to a request. It encompasses the trip to the server, the processing time of the call, and the return trip. Response time is a more valuable metric because it provides a broader view of the time it takes to get a response from your API.
A good response time is less than one second, and a great one is in the low hundreds of milliseconds.
The primary impact of response time is the perception that your API consumer has of your API. If your response time is greater than two seconds, chances are your API consumer will start to feel the delay. What’s more, it may stop them from using your API. Hence the importance of API testing.
Methods to improve API response time include:
- Use caching to reduce the time it takes to retrieve data from the server. Caching can be client-side or server-side.
- Optimize your API code to reduce processing time.
- Minimize the number of API requests by using batch processing or pagination.
- Use a content delivery network (CDN) to reduce latency for users in different geographic locations.
- Monitor your API performance regularly to identify and fix issues that impact response time.
Error rate is the percentage of requests that result in errors — the ones that won’t return a 200 response.
This metric is crucial for identifying issues with the API, such as bugs or compatibility issues. While it’s hard to give a gold standard of the percentage or error rate you should have, you can drastically reduce them by:
- Understand where the errors come from.
Here, API logs are your friend. A tool like Blobr can help you track such errors: you can filter calls by user and product and export them to analyze at scale how your API consumers use (or misuse) your API. If errors result from Bad Requests — then, your API doc may be misleading. And if there are lots of 429 Too Many Requests, check that your rate limit policy is clear for your users. A 500-style error? Then, the problem’s on your side.
- Implement clear error messages.
If your API returns only a 4XX code, chances are your API consumer won’t instantly understand what’s wrong with their call, and chances are they will replicate the error. A clear error message on the call and on your API documentation will help your API consumer implement your API and limit errors.
- Communicate with your API consumer — especially if you are versioning your API.
If API logs are very useful to see in real-time what your API consumers are doing, they are also a great tool for getting in touch with them. It’s especially true if you are versioning your API and deprecating parts of it. Sometimes, a changelog on your API doc won’t be enough to avoid breaking changes. This is why tools like Blobr can do the work for you and send automatic emails to users impacted by the changes.
Uptime is the percentage of time an API is available and functioning correctly.
This metric is essential in ensuring that APIs are reliable and accessible to users. Downtime can result in lost revenue, decreased user trust, and a negative brand reputation. Most companies have an SLA (Service Level Agreement) ensuring their customers that the API will be available for at least XX% of the time. Usually, it will be in the 99%+ range.
And don’t forget the difference between 99% availability — meaning several days of cumulated downtime per year —, and 99.9%, which cuts downtime to mere minutes per year. Downtime does occur, so make sure you’re not trapped by a picky SLA.
Free tools, like this one from Atlassian, provide Status pages displaying your uptime.
Business API Metrics
On the business side, API metrics help organizations measure the impact of their APIs on their bottom line. These API metrics are essential to analyze how to drive more revenue, engage customers and improve overall business outcomes. Some of the key business metrics and KPIs include:
Number of visits and Signups
Time to First Call alone won’t help you understand the possible sources of growth for your API business. For that, you need to have a closer look at the sources of acquisition of your API consumers and at the path they take before signing up.
- From your website to your API portal, with Google Analytics
Connecting your API portal to your Google Analytics is one way to have a unified look at your acquisition funnel. From your website to your portal, you can identify the best-performing referrers and sources of acquisition.
If you are advertising your API on Google, LinkedIn, or elsewhere, having this bird-eye view will help you double-down or cut campaigns based on how they perform.
If you bet on SEO, you will be able to see which content is working best and where you can improve (with additional tools like the Google Search Console).
And if you like Outbound, track down how your prospects are interacting with your message and landing pages with UTMs.
- Inside the API Portal or Store
Tracking how your visitors become API consumers is as equally important.
Some tools offer Page flows that help you identify patterns between visitors who convert and those who don’t.
It can help you improve the order of your API Products on your Product Catalog and how you introduce them. Or, for example, if many visitors check the pricing, but don’t go further, you may have a problem with your pricing.
Time to First Call
This one could have been on the technical side since it measures the time it takes for an API consumer to make their first successful call with your API. This is an essential KPI to track since it encompasses the whole acquisition and activation funnel.
From the first visit to your website to the subscription and subsequent first call, the TTFC can act as a good north-star metric for any API business.
But it will take more to understand and improve your TTFC. A better marketing message can be as important as providing better API documentation, for example.
For Twilio, a satisfying Time to First Call is less than 5 minutes. It works because it combines:
- A clear marketing message with identified use-cases → “One API for global, trusted messaging.”
- Top-notch onboarding → “Send your first SMS in minutes with this code.”
- Crystal-clear API documentation.
In short, Twilio explains why a productized API on a sleek API portal or store providing a unified place to discover, subscribe and test your API is so important to improve this key indicator.
The Adoption rate is the percentage of users who have adopted an API and actually use it.
This is a metric that will complete the Time to First Call, since it will measure the rate of API consumers who passed the testing stage and implemented your API in production. This is a crucial step in terms of measuring the success of an API, user engagement and adoption.
Keith Casey breaks this part into three phases: Exploration, Integration and Adoption.
- Exploration is just passed the First call. The API consumer now knows you and is much more likely to really start using your API. But, other priorities may delay or even stop them in the integration process.
- Integration is the next step and another chance for you to make a success of your API. Keith Casey identifies missing features and mediocre documentation as potential breakpoints, making API usage drop within a few days.
- Adoption is the final step. The implemented API starts to be used in earnest. This is when the API usage really starts to spike.
The interlude between the first two phases is the critical part, as it can go on for months. Ways to shorten this delay and speed up adoption are, once again, a neat first experience with your API, and the manner in which its productized versions respond to a requirement.
API Usage Growth & Monthly Active Users
This set of metrics is here to help you assess how use your API. There are two ways to exploit these metrics, which depend on your API and your business goals.
API Usage Growth is a look at the number of calls made. It can be an important KPI if you use Pay-as-You-Go pricing. Note that seasonality is sometimes very important and you risk experiencing quite the roller coaster — sometimes with 50%+ drops or increases between two periods.
Monthly Active User is more important if you use subscription pricing. It is also a key metric for gaining a more stable view of your growth and calculating your churn rate. Once again, this number alone won’t be that much of a help: one MAU making 10 calls and one making 10 million calls will not have the same value.
This is why you can combine those KPIs to have a clearer view. For example, you can take the average number of calls per MAU and use this metric to deepen the analysis. What is the average number of calls made per MAU who churned? This can help you establish patterns and take action to reduce or enhance your business strategies.
Don’t forget to use cohorts and predetermined periods of time to have a clear idea of your growth over a period. At Blobr, we use 2-week periods to analyze that very same data. We use the time to take an objective look at those numbers and to see what’s working or what isn’t.
This is the revenue generated by your APIs.
Once again, it is important to break down the overall number to gain valuable insights.
Which API consumer generated the most revenue? Which API Product is working best? Which referrer or acquisition channel is the most valuable?
Depending on how to crunch the data, all of those questions can be answered. And provide insightful data for the following teams:
- For the Marketing team, break down the value generated with the origin of API consumers and get a precise view of the advertising campaign's ROI, or which segment reveals itself to be the most valuable.
- For the Sales and CSM teams, see which users are generating revenue, and for those not using the API Products to their full capacity, how you can implement upselling tactics. With the Product team, try to implement Free trials or pricing strategies aimed at improving the revenue generated per user.
- For the Product team, see which API product is working best and if the pricing is adapted. How the endpoints are used and the way your API consumers are using your API Products can help you find new ways to productize your API.
Retention and Customer Satisfaction
Customer satisfaction is a metric that measures the satisfaction of users of an API. Good customer satisfaction means a better Retention rate.
Data is the first step in building better customer satisfaction and increasing retention. User Usage will provide you with your API’s consumer trends. With this tool, you can see if someone is making more or less calls and, more importantly, if they had to face more or fewer errors. If the number of calls is decreasing and the number of errors is growing, they can be the first signs of a potential churn.
Building a lasting relationship means anticipating these issues. We saw earlier ways to limit the impact of errors and communicate with your API consumer about breaking changes and new versions. API Logs are your best tool to prevent churn.
And if they aren’t always sufficient, keeping in contact can also help: a chat integration on your API portal or store enables a closer relationship with your customers.
If churn is unavoidable, those metrics and features can prevent it.
API metrics and KPIs are essential in ensuring that APIs are performing optimally and contributing to the success of an organization. They concern every team, from Marketing to Developers, by providing valuable insight.
The one thing you need to remember is that APIs are not just technical components, but also business enablers. Therefore, measuring the right metrics and KPIs is crucial for both the technical and business sides. Blobr provides those API metrics to help you succeed in your API business endeavors.