Why Google Is Eliminating Its Autocomplete API
Google’s Autocomplete function has become so ubiquitous, as users we barely notice it happening anymore. You start typing in a phrase in Google’s search bar, and it comes up with a handful of possible phrase completions—one of which is undoubtedly linked to your intention.
When this feature first debuted, people thought it was some kind of sorcery. Now, they take it as a given. But Autocomplete has had a wider impact than just on searchers—for years, Google has offered a direct API to its Autocomplete algorithm, giving developers the chance to find practical applications for the function. Unfortunately, this is no longer the case. Effective August 10, Google has pulled the Autocomplete API from availability, cutting off developers from the code.
How Developers Have Used Autocomplete
At first glance, it wouldn’t seem like Autocomplete could be used for much. However, developers have found a number of uses—both practical and ridiculous—for the API. For example, some developers have created games where users can guess what results Autocomplete would come up with for a given phrase (though technically, you can play this game yourself directly on Google). Other sites have used Autocomplete in the context of their own websites, offering a kind of in-app recommended search tool.
As a more important example for the SEO community, the Autocomplete API has been used by many keyword tools, including Ubersuggest. Using Ubersuggest, users can plug in a keyword or phrase and immediately populate hundreds—if not thousands—of related words and phrases that can then be used for content marketing or SEO campaigns. Because the information comes from Google’s Autocomplete in real-time, search marketers could confidently attest to the popularity of these recommended searches. Once that API is removed, these tools will all go defunct.
What This Means for the Search Community
The biggest impact this change will have is on keyword recommendation tools, which have almost exclusively relied on the Autocomplete API for years. Generating potentially viable keywords and phrases from scratch is going to be a much more tedious process, as you’ll have to either come up with them on your own or type your words into Google randomly and iteratively to collect the recommended search completions yourself one by one. Neither of these options is appealing.
Fortunately, Google still offers plenty of information on keyword search volumes and competition rankings in its own Keyword Planning tool within Google AdWords. Experienced search marketers should have no major problem finding an alternative solution to take the place of these Autocomplete-dependent solutions. Still, they were valuable while they lasted, and their results were highly valuable for thousands of marketers.
First, let’s take a look at Google’s official stance on the matter. In a recent blog post announcing the removal of the APl from public availability, Google insists that the reason is simple. The Autocomplete feature is intimately tied to online search—and while the team can feasibly imagine a handful of potential uses for the API beyond search, search is where it can and should live exclusively. As a result, the Google team wants to keep Autocomplete doing what it does best.
This answer makes sense, but it also reeks of corporate vagueness. If you read between the lines, it’s much clearer why Google would pull the API altogether. Google didn’t like the fact that other sites were using the API to generate their own keyword recommendations. Keywords, as an SEO strategy, are all but dead, but Google is still fighting back against their use. Google believes that anybody hunting for specific keywords to use in a campaign probably doesn’t have user experience as their main priority—and I have to say, Google is probably correct. Ubersuggest and similar suggestion platforms are valuable for brainstorming, but their entire existence supports the idea that keywords are still relevant for search rankings, which leads to worse content and poorer user experiences.
Seeing that Autocomplete is a low-key API (there aren’t many practical functions for it outside of keyword suggestions), Google’s decision to pull it from availability is warranted. Still, like many other search marketers, I’m going to missUbersuggest and its competitors.
Google’s Fine Line Between Open- and Closed-Source Systems
When it comes to company information, Google is all or nothing. It will fight tooth and nail against having to reveal the inner workings of its primary search algorithm, but it’s more than happy to share all the data about your site with you for free in Google Analytics. Google also offers more than 80 different APIs for various services, and attests to great things being done with many of them—as an example, Google often mentions how engineers use Google Maps data for their own applications.
So what’s the deal with this? On one hand, it seems like Google is all about making everything publicly available, but on the other hand, it seems like a top-secret corporate machine. But the truth is, there isn’t a real discrepancy here. It’s not about whether Google wants its information to be publicly available or not; it’s about whether or not something will actively improve the average user experience online. If Google believes an API will improve the Internet and user experience as a result (like with its Google Maps API), it will make it available. If it believes an API will only make things worse (like with Autocomplete), it’s gone. Like with SEO in general, everything comes down to user experience.
What can we help you with?
- Link building services for my company.
- White label link building for my clients.
- Major media brand mentions
- Something else (get in touch!)
Looking to grow your traffic?
Our managed SEO and social campaigns and high domain authority link building will increase your presence and organic search engine traffic.Request a rate card
Want more great resources?
Check out our new Resource Library, with over 100 expert articles spanning all aspects of online marketing, divided into 16 chapters.See our Resource Library