35 Domain Authority and ChatGPT Statistics for Software (2026)

If you run a software company, your monthly search report almost certainly opens with an authority score. Moz calls its version Domain Authority and scores it from 1 to 100;¹ DataForSEO’s runs from 0 to 1,000.²

Each tool computes its score its own way, but all of them are built mostly from the links pointing at your site, and Moz describes its score as a prediction of how likely a site is to rank in search results.¹

Link building gets bought against that number. So when someone at the leadership meeting asks why ChatGPT keeps naming a rival when buyers ask about your category, the score is the explanation on offer: they have more authority, so we need more links.

We set the score, and the pages ChatGPT reads before it answers, against the software brands it named in September.

The backlink score tracked who got named weakly. How many of the cited best-of lists mentioned a brand tracked it far more closely, and among the bigger brands the score hardly told them apart at all. That makes the score a poor number to run an AI-search channel on, and the lists a better one.

One limit before the numbers: this is a pattern among brands ChatGPT already names, in one month of answers. It says nothing about how a company gets named for the first time, and it does not prove that getting onto lists causes the naming.

Cited Lists Tracked ChatGPT’s Picks Better Than the Backlink Score

The answers come from our research, September 2026: ChatGPT, with search on, answering the buying questions of eleven software categories, from CRM and payroll to help desks and password managers, asked from the US, the UK, Canada and Australia.

Named means the answer’s text names the brand or one of its products. List presence means the share of the best-of pages the answers cited that mention the brand anywhere on the page.

The backlink score here is DataForSEO’s rank, its counterpart to Ahrefs’ Domain Rating.

Inside a single category the pattern is easier to see, because the brands are competing for the same answers.

Across every brand named in CRM answers or mentioned on CRM lists, the backlink score showed no relation to who got named; the lists did.

The lists do not settle everything: monday.com was mentioned on 43.20% of the cited CRM lists and named in only 13.04% of the answers. A mention on a page is not a recommendation on it, and this measure counts both.

The same held inside each brand’s own category, across all eleven, which is the fairer test: a brand in a crowded category sits on more lists and gets named more simply because there are more questions.

List presence and naming were measured on the same answers, and an answer may cite pages naming brands it would have named anyway. Read list presence as the better guide, not a proven lever.

The Score Measures Links to You; the Answer Is Built From Lists

The reason sits in how the answer is built. When a buyer asks ChatGPT for a CRM that suits a small sales team, ChatGPT rewrites the question into one or more searches, reads what comes back, and may cite the pages it used.⁴

Google describes the same mechanism from its side: AI Mode, its chat-style search answers, may fan a question out into many related searches before it answers, and Google says no special optimization is required to appear there.⁵

For a buying question, much of what comes back is other people’s pages. In a University of Toronto study, 93.5% of the sources ChatGPT returned for questions about well-known brands were third-party sites and 6.5% the brands’ own; for niche brands, 95.1% and 4.9%.³ Those were consumer brands, not software, so read it as the direction rather than the size.

In our answers, about a quarter of the cited lists (26.97%) sat on the site of one of the brands the answers named: a software company comparing itself with its rivals.

The rest were spread thin: the largest single publisher, G2, held 5.87% of the cited lists, and 82.78% of the sites behind them had one or two.

So the score and the answer measure different things. A backlink score is a fact about your own site. Whether your name is on the pages the answer reads is a fact about everyone else’s.

Brands on the Most Lists Were Named Most Often

Sort the brands into four equal groups, once by list presence and once by the backlink score, and see how often each group’s typical brand was named.

At the top, the list sort’s median brand was named in 8.89% of the answers and the backlink sort’s in 5.56%; at the bottom they barely differ (2.13% and 1.94%).

List presence is noisy among brands on few lists, and its second quarter dips below its first. It earns its keep at the top, where the shortlists are decided.

Among the higher-authority half of the brands, the backlink score hardly told one brand from another in ChatGPT’s answers; how many of the cited lists mentioned them still did.

A link-building proposal for a software company whose rivals already have more links should name the cited lists the work will put it on. Otherwise it is a plan for Google rankings.

Asked on Google AI Mode, the same recommendation questions came back in the same pattern, a check on ChatGPT’s finding rather than a study of Google’s own, since the brands and the lists came from ChatGPT’s answers.

Buyers Shortlist While They Compare, Which Is Where the Lists Sit

The surveys put these answers in the middle of the decision.

In 6sense’s own 2025 study of nearly 4,000 business-to-business (B2B) buyers, 94% used AI models to research what to buy, and their use peaked in the middle of the journey, most often to compare vendors’ offerings.⁶ That is the moment a buyer asks which CRM suits a team like theirs, and the moment the answer reads the lists.

The same study found the winning vendor was already on the buyer’s Day One shortlist, the list they hold before talking to any vendor, 95% of the time.⁶ If the answer a buyer reads mid-comparison leaves you out, you may miss the shortlist that decides the deal; the surveys cannot say how often that happens.

In Responsive’s own July 2025 survey of buyers at companies of 500 people and more, 56% of technology and software buyers named AI chatbots a top source for discovering new vendors, against 32% of buyers overall; 80% of them used chatbots at least as much as search, and 37% of all buyers said AI had helped them find vendors they would not otherwise have considered.⁷

So the cited lists act as a sales channel whose contents other people decide. Like any channel, it needs an owner, a territory, and a number.

What a Software Company’s Search Report Could Lead With

The owner is whoever owns pipeline from search today, often the person who manages the SEO agency. The territory is your category’s buying questions, one list of questions per category you sell in, because the comparison is cleanest inside a category.

The number is your share of the category’s cited lists that mention you, read beside the same share for the rivals ChatGPT names most, with the backlink score still in the report, one line lower.

These numbers cannot come from someone typing the question into ChatGPT once a month. Ask the same question again and the companies change: in 96.11% of the questions the repeat askings named a different set, and of the companies named for a question, 42.83% came back every time.

Your share of the lists only settles across many answers to the same questions, asked on a schedule.

Moving the share is work on other people’s pages: getting onto the lists your category’s answers cite, each through the route its author uses to choose products (who writes the cited lists and how each one picks), and onto the pages other companies keep (building mentions for software brands).

The reading under those numbers is the same check on every cited page, in every category you sell in, every quarter. It belongs to software, and the call on which lists the next contract buys stays with your search owner.