Surfacing new search terms after AMP launch
Every website wants to increase qualified traffic. More visits from more visitors is a basic goal of most marketing, whether it’s driving greater sales or additional ad views or just about any other metric. And what’s a prime method to earn more long-term, stable traffic? Improving your organic search rank and reach.
When we first started building Accelerated Mobile Pages (AMP), we knew they were lightning fast. Now, after substantial experience getting AMPs running for clients, we’ve noticed another, less-expected benefit along with that speed: greater organic reach. Implementing AMP has often expanded a site into new search query territory, generating additional impressions for more terms. And these new terms offer valuable opportunities to turn a one-time browser into a lifelong customer.
Growth in the unique search queries that put your site in front of new eyes is always worth appreciating, even if the gains are modest. But for one client, we measured a jump from 25K queries with impressions before AMP up to more than 58K queries after AMP (And that’s not even counting the partially overlapping 22K from the separate canonical property.) The comparison time periods were separated by only a few days, and the set of products was nearly identical. In another case, with similarly close time windows and stable products, we saw the query count go from 93K before AMP to 135K for AMP only after launch. More than 10,000 clicks were generated by AMP queries that weren’t present in the data before.
We don’t necessarily expect a near-doubling of queries before-to-after. So many factors influence search patterns outside of AMP, and our customizing to each client’s unique needs make it tough to predict a particular gain. What’s more, we’ve observed AMP contributing to an overall increase in total mobile-search terms, even when the query count for the AMP property itself wasn’t larger than the pre-AMP canonical. Recognizing the shift in total mobile queries that surface in search can require the careful combination of datasets from different properties. Nonetheless, we’ve consistently seen AMP generate search impressions for keywords that didn’t surface previously.
AMPs positively affect non-branded search terms
In terms of query types, much of the expansion we’ve noticed comes from non-branded and long-tail search terms. Individually, these are typically a small portion of traffic, but collectively they can have a huge impact, and can be more tightly focused on what your site offers. If you sell widgets, it’s unlikely you’d see “Widgets” or “Widgets for sale” as a new query, but instead more searches akin to “2018 shiny purple things-like-widgets.”
In terms of sales potential, non-branded, long-tail queries are a major way to drive new business. Visitors searching for your brand know about you and why you’re the best choice; these more general queries are how you capture a new potential customer.
How to look for query growth
You have a few different options if you want to look for these changes yourself. Each has strengths and drawbacks, so the best choice will depend on the particulars of your situation.
Google Analytics can expose the full query list during the current Search Console limit of 90 days in the Acquisition -> Search Console -> Queries report. This enables you to export the data and/or examine queries alongside other dimensions and metrics of interest. However, only one Search Console property can currently be linked to each GA property, and Search Console Property Sets aren’t supported. So, if your implementation uses multiple subdomains reporting to the same GA property (as with the Client ID API), you’ll only be able to see organic query data for one subdomain.
Google Data Studio was recently made broadly available, and it allows you to create connections to different data sources for reporting. It uses API hooks to pull data and can connect to both single Search Console properties and Property Sets. After connecting to the relevant properties, you can create a table displaying queries for the different data connections, with the total number of queries displayed at the bottom.
However, note that although properties have a verbose name (http://domain.com) in the data connection setup, property sets are designated as a hash value starting with “sc-set:”. This can make it somewhat difficult to know exactly which to choose (especially if you have several). You’ll want to load the property set in Search Console, then look in the URL for “sc-set:”. But this should only be a one-time headache, and it could make a lot of sense to present this info within the context of DataStudio’s visualization and sharing ease.
The Google Search Console API provides direct access to the data behind dashboard reports, including more than the maximum rows (currently 999) shown in the Search Console web interface. However, you’ll want a solid understanding of a scripting language (e.g. Java, Python, R, Google App Script) to write your own functions that make the proper calls. Alternatively, if you’re already working with Google Sheets, you’ll want to check out Search Analytics for Sheets, which provides an intuitive interface to data via the API. You’ll still need to calculate counts and other metrics from the full tables of queries, and your Sheets performance may degrade if you’re lucky enough to have a sufficiently large list of queries. Like Google Data Studio, choosing an existing property set will also require knowing its hash identifier.
Implementing AMP for your site can have many positive impacts on your bottom line. Growth of your unique query set is a huge way to get in front of new visitors, potentially converting them to customers. AMP promises ever-richer and faster user experiences. Whatever direction that AMP takes, WompMobile will continue its commitment to client success and the attention to detail that position us to build and support the fastest mobile websites in the world.