Founders ask me all the time: "Can your tool show me Reddit posts with buy intent and automatically reply for me?" My product can do the first half if you want it to. The second half is a clear no. Here's why: There's a whole tool category right now promising customers from Reddit "on autopilot". AI finds the threads, writes the replies, mentions your product, sometimes posts from accounts that look like regular users. What actually happens: Reddit now catches around 25,000 spammy posts and comments per day, using its own LLMs. Your vendor's AI vs Reddit's AI, on Reddit's infrastructure... a bad bet. And the business model is these tools work for a few weeks while the accounts are fresh, then the bans and call-out threads arrive, the customer churns, and the vendor needs a fresh supply of startups that haven't been burned yet. From my experience: Three years ago, our initial pitch for Octolens was a lead gen angle. We never ventured into auto-posting, but still, it attracted people who wanted the outcome without the work. And when leads didn't fall out automatically, they left. The promise was damaging our own signups. What actually works: The long game (sorry!). It's being active in threads not all about buying intent. Being useful, giving advice, placing your brand when it makes sense. Answering questions or clarifying misconceptions about your brand or competitors. And most of all: be honest about who you are, use your voice, and post yourself. All this helps with distribution: - It gets your brand surfaced in Reddit search, Google search, and AI search. - It improves your brand perception and credibility - You learn about what users want from your product or content And yes, occasionally you and I can get lucky: someone asks outright for a social listening tool, I pick it up with Octolens, I answer, customer converts. Only chasing leads on auto-pilot on Reddit, from what I can see, is long-term damaging for your brand.

September at a glance @Octolens: - MRR +13%, sixth month in a row of double-digit growth - API and MCP calls 5x compared to August - Companies building on our API +50% - Instagram and reviews went live as data sources While our users automate more and more through our MCP and API, I spent much of the month on good old things that don't scale. I qualified every inbound signup, helped fix their workspace setups, and got as many as possible on a call. Once the flow worked, I handed 90% of it to Claude. And to stay sane: a daily morning walk with my favorite little person 🌅

We deleted 200 blog posts end of last year and since then our Google traffic more than tripled. Last year we published hundreds of AI-written SEO pages. Exactly zero hit Google's top 10. This year we focused on quality. And search just became our #1 acquisition channel: 15% of new signups in spring, 20% in July, 25% in August, 36% (so far) in September. Google traffic went from ~1K to nearly ~4K monthly visitors. Here's the playbook we follow now: 1. Delete first. We killed ~200 thin AI pages with low quality backlinks and kept a few dozen worth maintaining. 2. Same AI tools, better inputs. The failed pages were written from models' imagination and little input. The new ones are written from data: live search volumes + SERPs from an SEO API (@ahrefs or @dataforseo), and verbatim quotes from real users on Reddit, Hacker News and X, every quote linked to its source using the @Octolens MCP. 3. Write the pages buyers search for. Best-of lists that rank competitors honestly with up-to-date pricing breakdowns. Alternative pages for tools people are leaving. 4. Customer stories for recognition and original content. Prisma, Vercel, Tally, PostHog, Render, Metabase. Original stories like these feed many other blogs and pages, which helps those rank too. Not to mention great content for socials. 5. Review everything. People try to sell you automated SEO that just posts blogs on auto pilot and makes your domain rating grow. But from my experience its rather damaging then helpful to your SEO and brand image. Taking the time to read over content and edit it is (at least now) still necessary. I recently open-sourced our blog writting process as a skill. github.com/octolens/seo-page…

"what b2b saas taught me about getting ghosted" just me and the time slot you picked

Add @octolens to mentionrank.lol The leaderboard you can't buy your way onto: rank = how often people mention you on 𝕏. No bids, no boosts.

Last year we published hundreds of AI-written SEO pages, and zero of them made it into Google's top ten. Now on the same domain, written with the help of the same AI tools, we rank number one for "social listening tools." A keyword where advertisers pay $11 for a single click, and the page brings in around 2,500 organic visits a month on its own. What we changed is the data the AI reads before it writes: - live search volumes and SERPs from an SEO API (@ahrefs or @dataforseo) - verbatim quotes from real users on Reddit, Hacker News and X, each one linked back to its source (@Octolens) In the video I run you through the steps.