This planner reads a list of queries and estimates which search features each one might trigger, based on the words in the query. It then says what a page needs to compete for that feature. You get a table, a count per feature and a CSV of the same rows.
It is deliberately not a SERP scraper. Fetching live results needs an API, a location and a lot of requests, and the result would still differ by country and device. Working from wording is weaker but honest: every guess shows the signal that produced it, so you can disagree with it.
What each feature is looking for
Nine features are checked. A featured snippet is suggested by a question opening, definition or process wording, or the word best. People also ask needs a question opening. A local pack needs local wording such as near me, opening hours or directions, or a service word next to a place. An image pack looks for images, photos, designs, ideas, wallpaper or template.
Video results look for video, watch, tutorial, walkthrough, demo or review. Product results look for price and purchase words. Top stories look for news and recency words. A knowledge panel or sitelinks needs navigational wording such as login or official, or a very short brand-like query. FAQ rich results need the words faq, frequently asked or questions.
A query that matches nothing is listed under standard organic results. That is not a prediction that the page has no features; it means the wording gave the rules nothing to work with.
A worked example
Ten queries produced 13 rows in the table: 11 named feature matches and 2 queries with no signal. The summary line counts all 13, so read the feature bars rather than the headline if you want the breakdown.
| Estimated feature | Queries matched | What it needs |
|---|---|---|
| Featured snippet | 4 of 10 | A direct answer in the first 40 to 60 words |
| People also ask | 3 of 10 | A visible question-and-answer section |
| Local pack | 2 of 10 | A verified Business Profile and a location page |
| Image pack | 1 of 10 | Original images with descriptive alt text |
| Top stories | 1 of 10 | Timely coverage with clear dates |
Three of the ten queries opened with a question word, so the tool reported 30% question queries and put featured snippets and People also ask at the top. how to check keyword density matched both: the question opening triggered the snippet, and the question wording triggered People also ask.
The result that matters most is the least impressive one. local seo services london matched the local pack twice over, through local and through a service word next to a place, and the table shows the first signal only. seo agency near me matched on near me. Together they make a thin but real case that two of the ten queries need a location page, not a guide.
Expect false positives and misses
plumber leeds is a local query to any reader, and it matched nothing, because the local rule wants near me-style wording or one of a handful of named cities. A town the list does not contain falls through to standard organic results.
At the other end, review suggests video results even when the searcher wants a written review, and ideas suggests an image pack even when the page is a list of text suggestions. The more specific the query, the more often the string rules will be wrong in one direction or the other.
None of this is fixed by adding more words to the list. A rule set grows until every query matches something, at which point it stops discriminating. The value is in seeing the signal and judging it, which is why the table has a column for it.
What the estimate leaves out
Features differ by country, language, device and even the time of day. A query can show a local pack in one city and none in another, and a featured snippet on mobile but not on desktop. The tool has no location and no device, so its estimate is a single guess for all of them.
It also does not know about AI Overviews, ads, shopping carousels, image packs that appear without visual wording, or any feature Google has added since the rule list was written. Those are absent rather than ruled out.
No click-through figures are given anywhere, on purpose. A CTR estimate per feature would look precise and would be invented.
Using the plan without over-trusting it
Run the list, then open the live results for the three or four queries that matter most. Search from the country you are targeting, on both a phone and a desktop, and note what actually appears. Where the tool and the results disagree, the results win.
Then look at what each feature needs and decide whether you can supply it. A video result needs a video, a local pack needs a verified profile and a location page, and a product result needs prices and structured data. If the asset does not exist, the feature is not a target, however likely it is.
For the page itself, pair this with the snippet simulator to see the title and description, and with the content optimiser to check that the answer sits near the top where a snippet would take it from.