Methodology

How these leaderboards are measured.

Everything on these pages comes from Rankxa, the tracker we built and run ourselves. This page states what we do, what we do not do, and where the method is weak, so you can decide how much weight the tables deserve.

What we do

The measurement, step by step

1

Collect the questions

Each leaderboard groups the prompts in our tracking set that ask a given commercial question. The number of prompts behind every table is printed on that page, because a table built on eleven prompts deserves less weight than one built on a thousand.

2

Run them repeatedly

Every prompt runs against ChatGPT, Claude, Gemini and Google AI Overviews on a schedule. Generative answers vary between identical asks, so a single run proves nothing and we never report one.

3

Extract the named brands

We record which companies each answer names and where in the answer they appear. A company mentioned in passing and one presented as the recommendation are both recorded, with position kept as a separate field rather than folded into a score.

4

Order and publish

Brands are ordered by how often and how early each engine named them, per engine. Nothing is blended across engines, because averaging four systems that agree 1% to 11% of the time produces a number that is true of none of them.

Where it is weak

Limitations we would rather state than be asked about

Every one of these is a real constraint on how far the tables can be pushed.

Answers are not deterministic

The same prompt returns different brands on different runs.

  • We report across many runs rather than one
  • Small samples move more, which is why the count is published
  • Treat a one place difference as noise, not as a ranking

Prompt wording changes the answer

Our phrasings are not the only ones a buyer could use.

  • Each leaderboard groups several phrasings of one question
  • It is still a sample of the ways people ask, not all of them
  • Wording effects are largest on broad, generic queries

Answers vary by person and place

Personalization and location shift what an engine returns.

  • Our runs are not signed in as any particular buyer
  • So these are a general picture, not your customer's exact screen
  • Local questions are the most affected

Being named is not being endorsed

We record that a company was named, not whether the answer was kind.

  • Position tells you prominence, not sentiment
  • A brand can be named as a warning and it still counts here
  • Sentiment is tracked separately in Rankxa and is not on these pages
The rule that matters

We are in these tables only where the data puts us

BusySeed publishes these leaderboards and BusySeed is a marketing agency, so the obvious question is whether we put ourselves in them. We do not, and there is no code path that could. On several of these questions we are not named by any engine at all, and those tables are published unchanged. A leaderboard a reader can verify in thirty seconds is worth publishing only for as long as it is reported straight, and it would be a poor trade to spend that for one flattering row.

Humans at the helm. AI in the engine.

The people behind the measurement.

Rankxa is ours. We built it, we run it, and the same team that reads these tables every month is the one that would work on your account. Different specialists, one accountable team.

Meet the team
Michael, Engineering Lead at BusySeedMichael · Engineering Lead
Amal, Automation Architect at BusySeedAmal · Automation Architect
Bryn, Solutions Engineer at BusySeedBryn · Solutions Engineer
The BusySeed Guarantee seal
The BusySeed Guarantee

We are serious about results.

If we can't increase your revenue by the 6 month mark, our team works for free until we do.

What clients say

What it is like to work with us.

★★★★★
“BusySeed created and managed our social media and wrote blogs that raised our SEO to a new level. We are getting new clients from organic search, and their lead generation campaigns made it easy to justify the investment by focusing on ROI.”
Leandro P., Engineer, Inviron
Leandro P.
Engineer, Inviron
★★★★★
“BusySeed demonstrated tremendous ability in understanding our needs and taking us by the hand in the process of designing a digital strategy from scratch.”
Alessandro Jarzynski, CEO, Tryger
Alessandro Jarzynski
CEO, Tryger
★★★★★
“Great professionals. They manage our company's social media and designed our website. I highly recommend them for social media management and web design.”
Raphael Costa, Owner, Meister Concrete
Raphael Costa
Owner, Meister Concrete

Want this run on your own category?

Same method, your buying questions, all four engines, plus a check on whether each engine can retrieve your site in the first place.

Get your AI visibility audit
Free strategy session

Want this measured for your own category?

Book a free 15-minute strategy session with founder Omar Jenblat. He'll look at your funnel live and show you exactly where the revenue is hiding. No deck, no fluff, no obligation.

Omar Jenblat, Founder & CEO of BusySeed
Omar JenblatFounder & CEO, BusySeed
  • A live 15-minute look at your funnel
  • No deck, no fluff, no obligation
  • Leave with 2-3 concrete growth moves

First, who are we meeting?

Three fields, then pick your time. We read up on you before the call so we open with something useful.

No sales sequence. If you never pick a time, we leave it there.