# DocsBot says better customers are barely offsetting a shrinking signup funnel DocsBot founder Aaron Edwards says higher ARPU and lifetime value have barely compensated for weaker signup volume, while the company’s 2026 attribution data shows a large gap between tracked AI referrals and self-reported AI discovery. The useful small-SaaS lesson is not that SEO is dead or AI search has won. DocsBot’s own numbers show how a channel can remain the largest share of conversions while the total funnel underneath it shrinks, and how 'Direct' can conceal the discovery path that actually influenced a sale. - Status: Active - Published: 2026-10-05T19:10:31+13:00 - Updated: 2026-10-05T19:10:31+13:00 - Categories: SaaS, Indie Business, Marketing & Distribution, SEO & Search, GEO & AI Search, Indie SaaS, Product & Growth, Conversion & Analytics - Tags: AI search, attribution, customer acquisition, DocsBot, SaaS growth, SEO - Canonical HTML: https://beyondthe.news/dossiers/docsbot-signups-search-ai-referrals-attribution-2026 ## What changed DocsBot founder Aaron Edwards published an analysis of first-touch paid-checkout sources from January through August in 2024, 2025 and 2026 alongside self-reported discovery data from demo bookings. He says signup volume has fallen in 2026, while higher average revenue per user and lifetime value have increased enough to cover the lost volume only narrowly. Search remained one of the largest measured checkout channels but fell from 48.5% of paid checkouts in 2024 to 40.6% in 2026. Recognizable AI referrals represented 4.7% of paid checkouts, while 18.2% of demo respondents said they heard about DocsBot through ChatGPT, Claude, Perplexity or similar tools. Word of mouth/referrals and Google/Search each represented 20.5% of demo responses. ## Why it matters This is a useful operating case for small SaaS because it shows three metrics telling different stories at the same time. Customer quality improved, measured search share remained high and revenue reached roughly $1 million ARR earlier in 2026, yet the founder still describes the top of funnel as weakening. The attribution gap also demonstrates why teams should not equate a browser referrer with the full discovery journey: AI assistants, recommendations and return visits can end up classified as Direct or another last-visible path. The data is one company’s founder-reported analysis and the checkout and demo populations are different, so it should not be generalized into a market-wide AI-search share. ## Search stayed large while the funnel got smaller DocsBot’s first-touch paid-checkout data shows Search declining from 48.5% of checkouts in 2024 to 40.6% in 2026, where it tied Direct. Edwards’ point is that share can remain impressive even as absolute signup volume falls. Ranking acquisition channels by percentage alone can therefore hide whether the total number of potential customers is shrinking. ## Tracked AI referrals and remembered AI discovery disagree Recognizable AI referrals account for 4.7% of paid checkouts in the PostHog data, mostly from ChatGPT. In a separate dataset, 18.2% of people who booked demos selected ChatGPT, Claude, Perplexity or similar tools when asked how they heard about DocsBot. Edwards explicitly warns against treating the second number as four-times-undermeasurement because the populations and measurement methods differ. ## Direct is becoming a bucket for missing context Direct rose from 35.4% of paid checkouts in 2024 to 40.6% in 2026. Some of that may represent brand awareness and repeat visits, but it can also absorb journeys where the initial influence was a recommendation, an AI conversation, a social post or another untracked touch. For DocsBot, Direct is therefore a measurement category rather than a clear acquisition strategy. ## Word of mouth matched Google among demo respondents In the self-reported demo data, word of mouth and referrals represented 20.5% of responses, exactly matching Google and Search. That makes customer recommendation at least as visible as the channels receiving more attention around generative-engine optimization, although the sample is demo bookers rather than all customers. ## Higher customer value is not a substitute for acquisition forever DocsBot crossed $1 million ARR in February 2026 after a pricing and positioning shift toward better-fit business customers. Edwards says higher ARPU and lifetime value have helped keep the business healthier as signups weakened, but only barely cover the gap. His next measurement goal is to join acquisition source to activation, payment, expansion and retention rather than stopping at checkout. ## Key details - DocsBot founder Aaron Edwards says 2026 signup volume has materially declined. - Search represented 48.5% of paid checkouts in 2024 and 40.6% in 2026. - Direct represented 40.6% of paid checkouts in 2026, up from 35.4% in 2024. - Recognizable AI referrals represented 4.7% of paid checkouts. - 18.2% of demo-booking respondents self-reported AI assistants as a discovery source. - Word of mouth/referrals and Google/Search each represented 20.5% of demo responses. - The founder reported DocsBot crossing $1 million ARR in February 2026 after pricing, onboarding and positioning changes. ## Builder takeaways - Track absolute acquisition volume as well as channel share; a channel can stay number one inside a shrinking funnel. - Keep tracked referral data separate from self-reported discovery instead of forcing them into one attribution percentage. - Treat Direct as 'unknown or unobservable path' unless you have evidence it is genuinely direct brand demand. - Join acquisition source to retention and expansion before deciding which channel produces the best customers. - Do not let improving ARPU or LTV hide a persistent decline in the number of qualified customers entering the funnel. ## What to watch - Whether DocsBot publishes absolute signup or checkout counts alongside future channel percentages. - The retention and expansion differences between Search, referrals, AI-influenced prospects and Direct traffic. - Whether the tracked-versus-self-reported AI discovery gap narrows as referral metadata improves. - Whether the company can restore signup growth without giving back the ARPU and LTV gains from moving upmarket. ## Uncertainties - The operating and revenue figures are founder-reported and not independently audited. - Paid checkouts and demo respondents are different populations, so their attribution percentages cannot be directly compared as one funnel. - The article does not publish absolute signup counts, making the magnitude of the acquisition decline impossible to calculate from the public data. - A first-touch referrer cannot identify untracked recommendation or conversational influence. ## Sources - [We Had Fewer Signups. Better Customers Barely Made Up for It.](https://uglyrobot.dev/articles/fewer-signups-better-customers) — Aaron Edwards · founder analysis · 2026-09-14T00:00:00+12:00. Primary source for 2024–2026 checkout-channel percentages, demo self-report data, signup decline and founder interpretation. - [We finally hit $1M ARR in February](https://uglyrobot.dev/articles/1m-arr) — Aaron Edwards · founder operating update · 2026-03-02T00:00:00+13:00. Context for DocsBot scale, prior flat growth, pricing/positioning changes and the company-reported $1M ARR milestone. - [Pricing](https://docsbot.ai/pricing) — DocsBot · official pricing. Current pricing and packaging context for the company’s move toward business customers.