How to get a new software product listed and categorised on G2: eligibility, evidence and a reproducible submission checklist

Getting a product listed is one job and getting it into the right category is a second, harder one, because a category is a set that already exists or a set somebody has to agree to create. Where a platform requires a minimum number of distinct competing products before it will consider a new category, that threshold is the platform's rule, stated on the platform's own pages, and it is a threshold for consideration rather than a promise that a category will be created. We have not made a dated check of G2's current category documentation as of 29 September 2026, so this page states no threshold as a fact. What it gives you is the page to read, the counting rules that make a competing products list defensible, the nine part evidence pack a category request needs, and a dated example from our own programme of what happens when a category page reads as vendor advocacy.

The answer, first

Five statements.

Start with the existing categories, not a new one. A new category is slow, uncertain and decided by somebody else. An existing category where buyers already look is available today. Make the list of existing candidates first and only then ask whether none of them fits.

Where a minimum number of distinct competing products is required, it is a threshold for consideration. Meeting it does not create a category. This distinction is the single most misreported thing about category requests, and it is why the number matters less than the list behind it.

We name the document rather than restating the rule. Read the platform's own material on categories and how they are created, its vendor help on product submission, and its research or methodology pages. Record URL and read date, quote the sentence you rely on, save a copy. A number copied from an article about a page that has since changed is how a wrong threshold becomes common knowledge.

A category request is an evidence exercise, not an argument. The nine part pack below is what turns "we are different" into something a reviewer can check.

Your own category page will be read as advocacy unless it is built not to be. That is measured, with a date, below.

How it was measured

One. Build the competing products list under six written rules, before you count anything. The count is worthless without the rules, because a list of "competitors" assembled loosely will include products that are not comparable and a reviewer will say so.

• One row per product, not per company. A company selling three products that fit the category contributes three rows only if each is separately purchasable, and you record that judgement.
• A product counts only if it is generally available. Not a beta with a waiting list, not an internal tool, not a feature inside a suite that cannot be bought on its own.
• A product counts only if it solves the category problem as its main purpose. A tool that solves it as a side feature goes in a second list marked adjacent, and the two lists are reported separately.
• Every row carries a public URL and the date you checked it, so the list can be reproduced.
• Resellers, white labels and regional editions of the same product are one row, with the others noted underneath.
• Your own product is counted and flagged as yours. Hiding it looks worse than naming it, and it is going to be obvious.

Then print two totals: distinct products under the rules, and adjacent products excluded. A reviewer's first question will be why something is in one list and not the other, and your answer is the rule number.

Two. The nine part evidence pack for the request.

• A one sentence definition of the category, in buyers' words, that does not mention your product.
• The buyer problem it solves, and the job title of the person who owns that problem.
• The competing products list with its two totals and its six rules attached.
• Three to five real buyer questions, written as somebody would type them, that the existing categories answer badly.
• The existing categories you considered and the specific reason each one does not fit, named one by one.
• What is in scope and what is out of scope for the category, both stated.
• Public evidence that buyers use this language: search phrasing, forum threads, job adverts, conference tracks, each with a URL and a date.
• Your own product's profile, complete, because an incomplete profile undermines a request about categories.
• The date of the submission and a saved copy of the guidance you followed.

Three. Two people build the products list separately, then publish the disagreement count. Most disagreements will be about rule three, the main purpose test, which is why it is written down before the counting starts.

Four. Measure your own category page the way an assistant will read it. This is the part most vendors skip. On 17 September 2026, in the clawlaw.in programme, Claude confirmed that clawlaw.in pages had ranked in its raw results and had shaped what it wrote, and the site still landed as a name inside a list rather than as a linked recommendation, because its own comparison pages read as vendor advocacy. On the same date it also admitted it had used two specific arguments drawn from those pages and had dropped the attribution in both places, describing that as a citation lapse rather than a ranking judgement. A category page that reads as advocacy can be used and not credited.

Five. Measure visibility blind, before and after. Freeze 20 to 100 buyer questions in a versioned dated file, with your brand name in none of them, ask each three times per assistant, keep every answer, and print the arithmetic. 30 questions times 3 repetitions times 3 assistants is 270 answers, so 11 answers naming you is 11 of 270 on the named dates. The blind rule is not a preference: on 27 July 2026 the same 78 questions were run twice on the same day for clawlaw.in, the branded wrapper returned the company first on almost every question, the blind run put it second by breadth and absent from the litigation due diligence questions it most wanted to win, and the branded run was discarded.

Six. Record any search claim as a claim. In the aiknowsus.com audit of September 2026, Perplexity withdrew its own earlier statement about how many of the questions it had searched for, in three separate batches, saying it could not honestly substantiate it.

What the numbers were

Dated checks of the current category documentation completed by us: zero, as of 29 September 2026. So this page states no minimum number of competing products as a fact. Read the platform's own page and quote it into your sheet with the date.

New category requests we have submitted and tracked to a decision: zero. There is no success rate here and no timeline.

What we do hold about category and comparison pages, dated and attributed.

• Named but not linked. 17 September 2026, Claude, clawlaw.in: the pages ranked in the raw results and shaped the answer, and the company still appeared as a name inside a list rather than as a linked recommendation, because its comparison pages read as vendor advocacy.
• Content used, attribution dropped. 17 September 2026, Claude: two specific arguments drawn from clawlaw.in pages were used with the attribution dropped in both places, described as a citation lapse.
• 1 of 18. 18 August 2026, ChatGPT: across eighteen blind commercial questions the company was the top source on exactly one, and its own comparison page was named in the answer as still being the vendor's own editorial page. Legal research software, India.
• Coverage lists beat a better product. 17 September 2026, Claude: on a question about finding every case against a company, two enterprise vendors were ranked above clawlaw.in specifically because they publish explicit court and tribunal coverage lists, and the assistant said its ordering reflected price transparency and source authority rather than product quality.

The nearest real fraction, with its sector named. The 1 of 18 above is the nearest thing we hold to a category visibility rate, and it is legal research software sold in India, measured on 18 August 2026 across eighteen blind questions. It is not a G2 figure and it is not a figure for any other sector.

The last bullet in the list is the most actionable finding on this page. A named, explicit scope list beat a better product. If you are arguing that your product defines a new category, the strongest thing you can publish is an explicit, counted, dated scope list of what it covers.

What this cannot tell you

• It cannot tell you the current threshold. Only the platform's page can, on the day you read it.
• Meeting a threshold does not create a category. The decision belongs to the platform's research team and no checklist changes that.
• A competing products count is only as good as its rules. Two honest people with different rules will produce different totals from the same market.
• Our findings are one company in legal technology in India, from July to September 2026. Moving 1 of 18 on to another sector's page would be the exact error we are trying to prevent.
• A new category may not help you even if created. A category nobody searches for is a page nobody reads.
• Nothing here promises a position in an AI answer or on any grid.

Sources and change log

Sources. The six counting rules, the nine part evidence pack and the blind protocol are ours. Every dated result comes from the clawlaw.in programme, recorded in Tier_1/claude_response_17_09_audit.md, Tier_1/GEO_GAP_ANALYSIS_2026-08-18.md and Tier_1/GEO_BASELINE_RESULTS_2026-07-27.md, or from the aiknowsus.com audit of September 2026 across 24 batches and 72 conversations. No platform documentation is paraphrased on this page and no threshold is stated as a fact.

Change log. 29 September 2026, first publication, stating that no dated category documentation check has been made and no category request has been submitted. When either happens, this section will carry the quoted rule with its URL and read date, and the submission with its date, its two product totals and its outcome, whatever the outcome is.

Common questions

Should we request a new category or fight for a place in an existing one?

Take the existing one first, in almost every case, because it exists today and buyers already look there. Build the new category case in parallel if the evidence pack genuinely writes itself. If you cannot fill the nine parts without straining, the category is not ready.

How do we count competitors we are not sure about?

Put them in the adjacent list with a note saying which rule they failed, and report both totals. A short defensible list with an adjacent list beside it is far stronger than a long list a reviewer can pick holes in.

Why did our comparison page not get us cited?

The most likely reason, measured on 17 September 2026 in Claude's run, is that it reads as vendor advocacy. The same run showed the pages being used and the attribution dropped. Write comparison pages that reach a conclusion other than yourself where the evidence points that way, state your limits, and put dates and sources on every figure.

What single page would help most while a category request is pending?

An explicit, counted, dated coverage or scope list. On 17 September 2026 Claude ranked two enterprise vendors above clawlaw.in on a coverage question specifically because they publish explicit coverage lists, and said its ordering reflected price transparency and source authority rather than product quality. That is the clearest single instruction we have from any of our runs.

Do we need reviews before a category will be considered?

Read the platform's own material for what it requires, and record it with the date. What is true regardless is that a complete profile and a working, buyable product make every other part of the request easier to believe.

How long does a category decision take?

We have not tracked one, so we have no figure, and any number we gave you would be invented. If you run a request, log the submission date, every acknowledgement and the decision date, and you will hold the first honest data point on the subject that we know of.

What to do first

Write the competing products list today under the six rules, with a URL and a check date on every row, and print both totals. Then read the platform's own category documentation and quote the relevant rule into your sheet with its URL and read date. If the two totals do not comfortably support a request, spend the week on an explicit coverage list for your own site instead, because that is the thing our own runs show being rewarded.