How to find where a bigger competitor is beatable
Every SEO tool will tell you a competitor is ahead. That is the part you already knew. The useful question is narrower and almost nobody answers it: given that they have years and links you cannot match this quarter, where does none of that help them?
That is a different question from "who is winning", and it has a different answer. Authority is bought with time. Some ranking surfaces are not sold that way at all, and those are the ones worth your afternoon.
Run it
docket attack https://yoursite.com https://theircompetitor.com \
--name "Their Brand" --demand
Docket crawls both sites, reads the link graph from Common Crawl, asks Google's public autocomplete what people actually search for in your subject, and returns the openings ranked by winnability — how little each one depends on authority you do not have.
What it says first
It opens by telling you the gap is real, because a tool that flatters you is no use. From a run against Screaming Frog, dated in the block below:
$ docket attack https://docketseo.app https://www.screamingfrog.co.uk/seo-spider/
# measured 2026-08-11
Link authority
you: docketseo.app was not found in the part of the graph scanned.
It is outside the top ranks rather than absent.
them: www.screamingfrog.co.uk/seo-spider ranks 2,661 of 117,963,409 domains
by harmonic centrality — the top 0.002% of the crawled web.
Their links are a real advantage and not one you close this quarter —
which is exactly why the openings below avoid competing on it.
Screaming Frog is genuinely one of the best crawlers ever written, it has been trusted for over a decade, and its link profile reflects that. If what you need is a desktop crawler for a large site, buy it. None of what follows is an argument that they are weak.
What it finds instead
Three kinds of opening, each measured on the two sites rather than inferred:
1. [strong] 1 search they answer and you do not
Real queries where they have a page and you have none:
"how to boost seo for free".
Why winnable: you are competing against a single URL, not their
whole site. Page-level quality is winnable; domain-level is not.
2. [strong] 9 real searches neither of you answers
From Google's public autocomplete, so these are queries people
actually type: free seo checker; where can i find keywords for
seo for free; what is the best free seo plugin for wordpress …
Why winnable: nobody has to be displaced on a query nobody has
answered. Their authority is irrelevant to it.
3. [worth taking] 4 of their pages contain no figures of their own
No percentages, sample sizes or counts — the copy asserts rather
than measures.
Why winnable: assistants quote specific claims. A site with
nothing quotable cannot be quoted, however authoritative.
Where the searches come from
Google's autocomplete endpoint publishes real queries, ordered by how commonly they are typed. No key, no account, no index. Docket infers your subject from your own pages, expands it through question and commercial framings, and then checks which of those questions your site answers, which theirs does, and which neither does.
The requests are capped, serialised and delayed between calls. It is somebody else's public endpoint and the run finishes in about a minute without registering as load anywhere.
What this deliberately does not do
Docket has no index of the web, and this feature is built so that it never needs to pretend otherwise.
- No rankings. It does not know where either site ranks for anything and will not guess. Every opening is a structural difference measured on the two sites.
- No search volumes. Autocomplete gives ordering, not counts. Docket reports the rank a query sat at and stops there — a monthly volume is a number it does not have, and printing one anyway is how this category of tool loses its credibility. Ahrefs and Semrush sell volumes; theirs come from clickstream panels Docket has no access to.
- No traffic or revenue projection. Docket projects its own score and nothing else.
- No completion that merely shares letters with your subject. Autocomplete matches substrings, so a site about SEO gets offered queries about Seoul. Those are dropped, and the run tells you how many it dropped rather than quietly shortening the list.
What to do with the output
Work the list top-down and stop when the winnability score falls off. The third kind of opening is usually the highest-scoring and the least obvious: publishing one page of numbers only you have — your own testing, your own customers, your own measurements — earns more citation in AI answers than ten pages summarising what everyone already wrote.
The second kind is the cheapest ground on the board and the easiest to act on this week. Write the page, put the question in the heading, answer it in the first paragraph.
Common questions
Can Docket tell me where a competitor ranks?
No. Docket has no index of the web, so it does not know either site's rankings and will not guess at them. Every opening it reports is a structural difference measured by crawling both sites — a rich result they cannot win, a crawler they have blocked, a question neither of you answers.
Where do the searches come from?
Google's public autocomplete, which publishes real queries ordered by how commonly they are typed. That ordering is a popularity signal, not a search volume. Docket reports the rank a query sat at and does not print a monthly volume, because it does not have one.
Does it need an API key or an account?
No. The link-authority comparison reads the Common Crawl hyperlink graph and the query discovery uses Google's public autocomplete endpoint. Neither needs a key. Both are capped and rate-limited so a run does not register as load.
Is this an argument that the competitor is bad?
No, and Docket opens by saying so. It reports the size of their link advantage first, in a number, precisely so the openings underneath are read as places their authority does not apply rather than as weaknesses. If their product fits your job better, buy theirs.