The New Twitter API Pricing Structure is an Expensive Failure of Their Ads Product

The new Twitter API pricing structure that was announced earlier today is incredibly expensive and will kill off most smaller apps. According to this tweet by the official TwitterDev account on February 9th, 2023 Twitter will begin charging developers to use their API for applications. The pricing for the Twitter API (as far as anyone knows at this moment) is located at this link.

TRIGGER WARNING:If you are a developer who has built an app on the Twitter API this pricing structure is going to give you minor heart palpitations.

Which of the Twitter APIs Will Cost Money?

Both versions of the Twitter API, including Twitter API 1.1 and Twitter API 2.0, will cost money to use.

Will This Stop Developers From Using the Twitter API?

Absolutely. Most Twitter apps use the free usage tier because realistically Twitter apps are not profitable. They are small scale social apps that are used for growing your main application’s user base or are just for fun. Even more serious Twitter apps or apps that use major data access for larger data mining are hardly profitable and rely on the extremely low cost or free access to the data to continue to service Twitter users on the platform.

Is There a Free Tier For Small Apps?

Yes… but it is only for Sandbox apps (meaning the applications are not live for most users) and the current request limit is 250 requests PER MONTH. The average development cycle of a small app even for an experienced developer will make hundreds of small API calls during the integration process to debug their integration and insure that it is working correctly for users. This includes not just the creation of the app but automated testing of the integration to make sure it stays working as other code around it changes. This will make it incredibly difficult to even create new Twitter apps as a developer.

Argument: The Servers Cost Money. Twitter is Losing Money. Twitter Needs to Charge Money

This is an incorrect argument based on a simplification of how a hyperscale social network like Twitter works as a business. Yes, the Twitter Platform is a major cost center for Twitter. However, the Twitter API (both 1.1 and 2.0) currently have many, many limitations on usage for apps that we previously much more permissive and allowed much larger tiers of free usage. Could Twitter have started charging between $10/mo and $100/mo based on usage for the API and it would have been OK? Most likely. But these new prices are so far out of line with the economic models of Twitter apps and of course are unlike the rest of their industry.

If you look at other competing platforms like Facebook, Instagram, or Google they all offer their APIs with some restrictions but they give extremely large amounts of usage and of course all their APIs are free to use. How do Twitter’s competitors keep offering their APIs for free with such large amounts of usage? Ads.

Twitter Charging Extremely High Prices for their API is a Failure of their Ads Product

It’s no secret the ads on Twitter are terrible. This is well known to anyone who has used Twitter extensively for many years as a normal person. On the other side of things, the advertisers who are making Twitter ads and paying for user clicks, have known for many years that most Twitter ads are not profitable and do not “back out” (this is a marketing industry term meaning the ads are a profitable way to spend money). Buying Twitter ads is rarely a high priority in any marketing department and is usually an afterthought. How does a failure of the Ads product result in Twitter needing to charge so much for their API?

Without an excellent ad product Twitter is not able to generate enough revenue to support the business like their competitors do. This results in the API being a highly successful product and growth strategy while being a major cost center that Twitter cannot pay for.

What is Elon Doing About Ads?

After Elon’s takeover of Twitter he had a major focus on the ad business because Twitter is bleeding cash horrendously. Unfortunately he lost a lot of advertisers due to the type of contentious content he hyped on the platform. He spent a lot of time in the Spaces he was in talking about the display ads and how to make them “relevant” to users hoping that there were secrets to magically making ads more relevant to users with technology.

Can Twitter Make Better Ads?

It may be too late for this. Elon lost many of the larger ad purchasers when he started treating Twitter as his personal playground rather than acting like a CEO running a failing business. This was a major blow to existing ad revenue stream. On top of that he was searching for answers to the “relevance” problem looking at things like machine learning or AI; unfortunately those avenues alone will not provide the substantial additional revenue. His primary focus was to show “products that people want to buy”; Elon was clearly hunting for the Instagram ad revenue model. He should be more focused on the Facebook style ad methodology that doesn’t have quite the focus on pushing to ecommerce

Elon also had the ads team try to surface more ads on the app. That may have increased revenue slightly but without excellent ad inventory to display that could actually make users MORE ad blind.

Are There Simple Solutions to the Twitter Ad Problem?

The easy solutions to improving ads and their relevance have so far been ignored. The only two simple solutions that Twitter could take overnight would be:

1. Push the text that says “Promoted” to the top of the ad unit (exactly like Facebook, Instagram, and Google do.) This seems counterintuitive at first: wouldn’t that REDUCE clicks on ads since users know they are ads? Yes and no. The problem with users not knowing something is an ad is that it incentivizes and optimizes for clickbait ads. Advertisers want you clicking on their ad before you know it is an ad. This makes advertising and advertisers on Twitter heavily focused on clickbait style ads rather than creating high value adds and trying to find relevance to users with quality ads that drive clicks. This is a very, very minor fix engineering-wise but it would also require shifts in the overall ad ecosystem (ad content & advertisers shifting) before it would pay off. This may not be possible given Twitter’s short runway.

2. Include the profile pictures and usernames of any users that interact with the add below the ad but above the tweet metrics. This includes any followers, following, 2nd degree followers or following, celebrities, or major brands that have liked, replied, or view the original tweet or who follow or engage with primary account that created the ad. This gives social pressure to the ads which would greatly increase their relevance to most users.

Kick You Are Such a Whiner, Give Us Big Solutions

Again, unfortunately it may be too late to save Twitter’s revenue via ads. There were two major plays Twitter needed to drive at using a combination of the the API + Platform + Ads team over the last years and they have failed to create the necessary products that would allow them to feature match with competitors and get the level of ad revenue they need.

Twitter Connect – The Missing Login Link

1. “Twitter Connect” There is no comprehensive web Javascript library similar to Facebook Connect for Twitter that works well. There is a JS lib that they supply which is mostly used for correctly rendering tweets but it isn’t pushed heavily as a login solution. If they had pushed more aggressively for using the JS lib as the standalone login library for applications this would have given them a better contextual understanding of what people are looking at *when they are NOT on Twitter* and that is a large part of Facebook’s strategy around understanding ad targeting. This would have also unlocked much more powerful ad campaigns particularly around retargeting which is a major focus of ecommerce advertising.

This would not be a huge technical effort for their Platform and API engineering team even after the Elon layoffs but unfortunately it relies on external website adoption. With the increase in the API costs and level of desperation to drive revenue it’s unlikely that even creating this JS lib would get the adoption to reach critical levels on external websites to heavily impact ad revenue.

“Twitter AdSense” – The Ad Platform That Never Was

2. “Twitter AdSense”. Twitter doesn’t get as much traffic as their competitors (Facebook, Instagram, Google Search, GMail) and so they can’t realistically supply the same level of ad inventory as their competitors. In order to drive more ads from their ads platform Twitter needs to be showing ads on external website to get the additional impressions. This is how the Google AdSense model works

You can see this model in action on this blog. I include a Javascript library in the header of this blog supplied by Google AdSense which automatically places ads in the spaces I tell it to. If you, dear reader, click the ads then the advertiser pays Google $1 and then Google pays me $0.20. If 1000 people view a blog post I also get about $0.02.

If Twitter had a similar Javascript library it would have been possible for some websites to adopt their ads on their website. Advertisers simply don’t have this option with Twitter. It is not a trivial thing for their Platform + API + Ads team to build and so with recent layoffs I do not think they would reach a level of product that would allow them to compete with Google AdSense.

Twitter is in a Bad Position, Hence the Twitter API Prices

As you can see from this blog post the new Twitter API prices are not Elon trying to get rid of bots. The new pricing does reflect Elon trying desperately to turn Twitter in to less of a money burning machine: it makes sense to charge for the API. But the actual level of the pricing, far too expensive or exorbitant for almost all of the apps to actually pay for, is actually a major failure of the overall situation of ads on Twitter and has nothing to do with the very successful Twitter API and Twitter Platform.

Don’t Blame the Twitter API or Twitter Platform Team

It isn’t their fault they made an incredibly successful product that may be the last viable source of new revenue for a desperate Twitter.

No AI or ChatGPT was used in the creation of the blog post. All spelling errors or grammatical mistakes are mine.