How Much is Your API Worth?

  October 09, 2015

What most people are willing to pay for it.

This might sound trite, but it’s true. The basic laws of valuation don’t go away just because we’re living in a world of Microservices, PaaS and Big Data on demand. It doesn’t matter if you are giving your API away, any more than it applies for trying to get rid of a litter of “free” puppies. If the user is not willing to make the investment of time, attention and support costs, then your API is not worth much. But, there are APIs out that are very valuable. So let’s take a look at some of them and then see if we can apply their success to determining the value of your APIs.

Do For Your Users What They Cannot Do for Themselves

Katrina Nicolet, a software developer in Bay Saint Louis, Mississippi cuts right to the chase. One of her first experiences coding using an API was with Google Maps. When asked why she found Google Maps valuable, she got right to the point, “Using that (Google Maps) versus writing your own mapping system is almost a no-brainer.”

Jay Drake, CTO of Chargebacks 911, a company that helps clients avoid retail chargebacks, makes a similar point. His company uses the set of APIs from Twilio to manage voice and text messaging. Says Drake, “...they have everything I could possibly want from a voice/text API. It's like having an Asterisk VOIP system available without the upfront costs and maintenance of setup.”

As I mentioned before, your API is worth what people are willing to pay for it. Companies are willing to pay money just as long as they get the services and data they need.

A Google maps mashup, for example, costs money in volume. The current rate for between 25,000 and 100,000 map load a day is 50 cents per 1,000 loads. Nicolet agrees the investment is work the price in high volume, commercial situations

Drake points out that his company pays for the Twilio services, and yes, cost is a consideration. Drake continues, “It is always a matter of weighing the costs, considering all of these factors - what the price will be to build and maintain something compared to the cost of using something available….”

Trizic, a publisher of tools for online finance management, pays for third party APIs. The company uses Datadog for gathering performance data, Zapier for application integration and Mandrill for email fulfillment. According to CTO Steve Mays, with regard to Mandrill, “There is no better way to send email.”

The important thing to notice about Mays’s statement is not that Trizic uses Mandrill, but that Mays thinks there is “no better way! You don’t want your users just to consume your APIs, you want your users to be fervent enthusiasts. An API that serves a community full of fervent enthusiasts is worth more than an API promoted by any one salesperson.

Make Your Community Wealthy

Speaking of community full of enthusiasts, SketchUp Ruby API allow  to create plugins that extend SketchUp’s 3D modeling capability on its desktop application. Using the Ruby API is free. Extension developers can distribute their plugins at SketchUp’s Extension Warehouse. The Extension Warehouse is similar to a typical App Store where extensions can be distributed freely, and SketchUp receives a percentage of the sale price.

Presently there are over 300 extensions available and they have had over 4 million installations. SketchUp’s thinking is to make the SketchUp community wealthier. To quote Jeremy Walker, a Software Engineering Manager at SketchUp at Trimble Navigation, “Yup, it's all about loving our users. By making our API free and publicly available, we encourage growth of our developer community and innovation by third parties that will only add more value for our users.  So although we aren't collecting revenue by selling our API and tightly controlling its use, we believe that giving away our API for free will ultimately result in more value for our users and we trust that this will be reflected in increased brand recognition and ultimately, increased sales of SketchUp Pro.”

Ease of use counts too. Your API is going to be worth more to a consumer if you provide complete, easy-to-understand documentation with examples that show exactly how to use each call. Again, Jay Drake, “If I have two APIs that are reasonably similar in costs otherwise, the good documentation is likely going to win.”

Ease of integration is important. Many companies will use a set of APIs as an underlying subset of to provide value-add to their offerings. A critical item is how easily the APIs can map into his underlying system. If a company can be a client of several APIs, ingest third party data into their own systems and build a brandable API layer on top of the APIs they consume, the value of each constituent API increases. The trick is ease of integration. Ease of data point mapping is essential.

How to Calculate the Worth of Your API

If an API is worth what someone will pay for it, then how do you price it or propose one internally? Getting a hard dollar amount on the worth of your API is hard. There are a lot of factors in play. Allow me to offer one concrete method of measuring API value.

Your API is worth the cost that each of your users would incur were each to have to create the service themselves divided by the number of competitors in your space + 1, dividing your user count by 10,000.

API Worth = ((Number of Users/10,000) x (Dev Hours to Create Service x Average Hourly Cost of Dev))/Competitors + 1

How does this play out?

OK, let’s say that I have an API with an imaginary service: www.mygraphic-service/thumbnail/transform/

The service takes an image file, jpg or png, and converts it to a simple thumbnail. Not a rocket science service, but one that might be useful. Let’s say that it took me 80 hours to create and publish it. Also say that if I were I to have hired someone to do the work, I’d have to pay $50 an hour. (I know, call me cheap.)

Next we publish the API. After some time we find 5,000 registered users. Also, let’s say that I did a cursory search and found out that there were 10 other services out there that made thumbnails. So here is how the math plays out:

 

V = ((5000/10000) x (80 x 50))/10 + 1

Thus, V = $181.

 

Here’s how I arrived at the calculation. First, my thinking is that a real world API should have at least 10,000 users. I arrived at this constant by intuitive hunch. (So much for advanced statistical analysis.) Sadly, my API that has only 5000 users is essentially half real world. If I had a million users, my API would be 100 times real world, 1,000,000/10,000 = 100.

The developer hours times hourly rate is the way I calculate the cost it would take one of my users to do it himself or herself. Obviously my little thumbnail service is easy and low cost as reflected by (80 x 50). So yes, a user could do it himself or herself. However, if we have a development cost number that looks like (5,000 x 50), the cost of 5 developers working half a year at $50 an hour, which equates to $250,000, things change.

Lastly, we have the competitor + 1 number. The plus one (+1) is my service. So, if I am the only person providing the service without any competitor, then I own the market, so to speak. Hence, divide by 1. However, the more competitors I have, the greater the opportunity for a user to jump ship either for cause or on whim. Also, when the user does go to a competition, he or she does not have to make the service himself or himself.  Thus, the benefit of my developer cost is diminished.

So, take another shot. Let’s say I improve my API to provide copyright registration based on jurisdiction and watermarking of a submitted image upon copyright. This is a much more valuable service. Let’s say that my service catches on and has 100,000 users worldwide and it turns out that I have 5 competitors. Lastly, let’s say it took 2 developers five hundred hours each (1000 hours) to bring the service to market.

So let’s run the numbers again.

V = ((100,000/10,000) x (1000 x 50))/ 5 + 1

V = $83,333

So, there you have perhaps the simplest way to value an API that could possibly work. Take what you like. Leave the rest.

Here are the Takeaways

  • Serve your community. Know the people and companies to which you will be providing API service. The more your serve our community, the more your API is worth.
  • Don’t feel as if you have to give it all away. If you provide value to your community, they will pay a fair price to use your API. Your API can be worth the price of admission, particularly if your user community considers you the best around.
  • The easier it is to understand, use and integrate your API into a user’s system, the more your API is worth to them.
  • If you need an informal way to calculate the worth of your API, use:

API Worth = ((Number of Users/10,000) x (Dev Hours to Create Service x Average Hourly Cost of Dev))/Competitors + 1

A Little Closing Parable

Allow me to share a parable.

A little over 7 years ago my wife and I took home a small, white, terrier dog. He was endearing, fun to walk and cute as all get out when he put his head on my lap while I watched TV. When left alone, he did not tear up the couch. When he needed to do his business before walk time, he went to the shower. I didn’t even have to train him. A few weeks into his stay with me, the rescue organization that I got the dog from asked for $250 to cover his kennel stay and shots. I paid it gladly. The dog met a need in our life, bringing us untold amounts of joy and entertainment.

The same can be said of an API. If your customers take it home and it's easy to use and provides them with what they need without requiring an inordinate amount of attention and support, they'll gladly pay a fair price. And, many times they'll become your best sales force.

How would you calculate the cost of an API?