Reading SERP Features (PAA, snippets, local pack, AI Overviews)

Full transcript

1,021 words · script draft, under review.

Intro

Two keywords. Same search volume. You rank third on both. One sends you plenty of clicks. The other sends almost none.

Same position, same demand, very different result. The difference is everything Google puts on the page that isn't a normal link.

A box that answers the question directly. A list of related questions. A map with three local businesses. An AI-written summary at the very top. On some pages, the first organic result starts halfway down the screen. On a phone, often below the fold.

In this episode we go through the main SERP features: what each one means for clicks, what it tells you about intent, and how to detect it in the API data from episode twenty-seven.

Featured snippets

A SERP feature is any element on the results page that isn't a standard organic link. We start with the featured snippet: a box near the top that quotes a short answer from one page, with a link to it.

What it means for intent: the query has a short, factual answer. A definition, a number, a list of steps.

What it means for clicks: it cuts both ways. If the snippet answers everything, many searchers never click anything. If the answer needs more context, the snippet page can get a large share of the clicks. And the page in the snippet is usually one that already ranks on page one, so it's a reachable target.

In the API data it shows up as its own item type, featured snippet, with the domain and URL of the quoted page. Record whether one exists, and who owns it.

People also ask

"People also ask" is the box of related questions that expand when you click them. Each answer comes from a page, with a link.

For clicks, the box itself takes space but sends relatively little traffic. Its real value is as research. The questions are Google's own list of what else people want to know about this topic. They're often perfect for headings or a short questions section on your page.

In the API it's the item type "people also ask", with the questions inside. Store the questions in their own table. Across a hundred keywords, the questions that keep coming back are the subtopics your page has to cover.

Local pack

The local pack is the map with usually three businesses under it. It appears when Google believes the searcher wants a place nearby.

That's a strong intent signal. If a keyword shows a local pack, the searcher probably wants to visit or call someone. A national online shop or an information site will have a hard time getting clicks there, because the most visible spot on the page belongs to businesses with a map listing.

In the API it's the item type "local pack". If you run local keywords, remember that location is everything. The same keyword in two cities gives two different packs. We'll look at that in episode thirty-four.

AI Overviews

AI Overviews are generated summaries at the top of the page, with links to the sources they draw on.

For clicks, they work a bit like a featured snippet, only bigger. Simple questions may be answered on the spot. But they also cite sources, and being one of those sources is a new kind of visibility.

For analysis, two things matter. Does an overview appear for this keyword, and is your site among the cited sources? In the API data, it appears as its own item type, AI overview, with its references. Check the documentation for your provider, because the way these are returned is still changing.

I treat this feature as something to measure, not something to predict. Track how often it shows up across your keyword set from month to month. That trend tells you more than any single snapshot.

The others, briefly

A few more, briefly. A video row means people want to watch, not read. An image row means they want to see. A shopping row means they're ready to buy, and it's a strong transactional signal. Top stories mean the query is about news, and freshness matters. A knowledge panel means Google sees the query as being about a known entity, like a brand or a person.

Each of them takes space from the organic results, and each one says something about intent.

And remember the device. The same set of features takes far more room on a phone. A shopping row, a question box and an AI summary can fill the whole first screen on mobile, while on desktop the first organic link is still visible. If most of your searchers are on phones, judge the features on the mobile results page. We compare devices properly in episode thirty-four.

One more thing to store: who owns the feature. A featured snippet or a video row that a competitor holds is a different situation from one that nobody in your niche has claimed.

Measuring features across a keyword set

Now combine it. From your stored snapshots, build one table: for each keyword, which features appear, and where the first organic result starts. You get that from the rank absolute of the first organic item.

Then summarise across the keyword set. Here are made-up numbers for illustration, for fifty keywords on example-shop.test. A question box on most of them. A shopping row on about a third. A video row on a few. And on some keywords, the first organic result doesn't start until position five or six.

Those last ones are the keywords where ranking third isn't worth what the volume suggests. That's the answer to the puzzle from the start of the episode. And when you compare this with your own Search Console data, features often explain a low click-through rate better than your title does. That's exactly why episode eleven checks the results page before rewriting a title.

Wrap-up

So: every feature takes clicks from the blue links and tells you something about intent. Detect them by item type, and track where the first organic result starts. The SERP feature cheat sheet is linked in the description. Next episode: we look at who ranks, and sort them into forums, brands, niche sites and aggregators. See you there.