Discovered Labs

SEO / AEO rocket fuel: This study organically acquired over 2,000 backlinks ($400k+ in retail value) and is growing exponentially, increasing by ~100 referring domains every month. (We produced this study for a B2B SaaS client of ours) If we took the average costs of building a backlink, the retail value of this study's link profile is somewhere between $400k - $600k. That's a lot of authority being pushed to commercial pages on this website. B2B SaaS companies are sitting on so much first-party data that nobody else has, it's the perfect ingredient for differentiated content that ranks and gets cited by LLMs. Here's the process we went through: (1) Topic and angle validation A mistake companies make is going after topics and queries that aren't grounded in demand. If nobody (incl. LLMs) discovers your study it won't succeed. These projects take dozens of hours to produce so we want the biggest probability of creating a splash. - we did an N-gram analysis of GSC to find demand patterns - we researched existing studies that were outliers - we found the sweet spot of demand <> narrative overlap (2) Data requests and analysis Once you have a better idea of the questions you want to answer in your study (and rough idea of queries to target), you pull the data. Some clients do this or our in-house engineers do it. This study analysed billions of data points - not something you can easily do natively ChatGPT or Claude without hallucinations. It requires real software. (3) Content production & design This is where we go through the editorial and SEO process. The questions we ask at this stage: - what queries do we want to target? - what's the overarching narrative? - how does this connect to our value props and differentiators? - what are the key insights for the reader & how might we visualise them? - are we optimising for discoverability and extraction? An asset like this can lift so many pages. Every internal link, anchor text, CTA and page component really matters. (4) Distribution Arguably the most important step that gets your study in front of your ICP. Many ways to do this: - PR / proactive link building - paid - influencer - owned social profiles (founder, employees, etc.) And don't forget to refresh all of your older, relevant content with the fresh data to experience a nice performance uplift :)

Man, does anyone else think AI Search / AEO advice sounds like witchcraft with hand-wavy tactics? It's like you've gotta add secret schema to your website that whispers to Sam Altman and in return he'll reward you with some citations Or clap your hands 3 times and say "llms.txt" in the mirror to get mentioned in Claude The advice on here can get a bit silly sometimes, so we made a short film about it called Unconventional Methods. (S/O to Ognjen Bošković and Sam Mackay at Artificial who brought this to life!) There are real edges in AEO - across clients we’re at 10,000s of conversions a month from LLMs - but without data telling you which ones work you may as well be picking your tactics off a tarot card 🔮 The team thinks we should do a follow-up episode on Jimmy’s story, what do you think? 😂 P.S. Discovered Labs is a team of AI/ML engineers, researchers and marketers who run end-to-end SEO/AEO services for companies like Granola, Instantly, Composio and other Series A+ SaaS companies. Click here to learn more: https://lnkd.in/eHVf23k2 *No witchcraft included

We ranked our client (SaaS, Series A) in AI overviews AND position 1 on Google and it pissed off their competitor 😬 A piece of content created by Discovered Labs was ranking in AI overviews and position 1 on Google within hours of being published, appearing higher than G2 and Gartner comparison pages. Mr. Competitor was not happy at all. Fortunately, the client didn't care. This is something I find funny about the SaaS industry: Competing in the SERP? You're evil and going to hell. But it's totally fine to: - Use software that scrapes your competitor's site every week to monitor changes (it's called product marketing!!) - Create BATTLE (⚔️) cards that enable sales reps with passive-aggresive talk tracks any time a competitor is mentioned - Runs "us vs them" ads, launch ABM campaigns that target competitor customers, etc. Apparently competitive marketing is ok as long as it's not in public view - has to be sneaky so nobody notices! Anyways, this piece of content was created using our CITABLE framework. It's based on our own research as well as research from Google and Anthropic. (DM me if you'd like the breakdown) Although we provide marketing services, we have a team of AI/ML engineers that are constantly researching and trying to develop edges for our clients. My co-founder Ben worked on transformers - an underlying tech for LLMs - and autonomous robots before we partnered. We've been able to do this (AIO, P1 on Google + mentions in LLM) fairly consistently for clients. -- P.S. Discovered Labs is a team of AI/ML engineers, researchers and marketers who run end-to-end SEO/AEO services for companies like Granola, Instantly, Composio and other Series A+ SaaS companies. Click here to learn more: https://lnkd.in/eHVf23k2

Just asked Ana-Maria, VP Marketing at GetVocal (Series A) what it’s been like to work with us on SEO and AEO, here’s what she said: This was a zero-to-one program. When we started the site was effectively blank with unresolved technical issues and because the product evolved fast, historical content and positioning had to be constantly refreshed. Due to low domain authority, the strategy was to capture as much commercial impact as we could from AI search in the short-term where things are less competitive, while building a strong presence on Google’s SERPs for long-term defensibility. Here's how we approached it: (1) Build a combined SEO and AEO plan grounded in real buyer language We sampled 500+ prompts on a regular interval across ChatGPT, Claude, Gemini and Google AI Overviews. We read the reasoning and sources each model used before answering to help us understand what’s having the biggest influence. This exposed gaps within topic clusters we wanted to win and informed a large part of the content roadmap. (2) Created and refreshed content against those gaps We published net-new pieces where we could cover a topic gap and refreshed existing content when positioning changed or individual pieces were underperforming. All content is planned and produced with our team of in-house editors and goes through our CITABLE framework to ensure it’s optimal for both Google and LLMs. (3) Put GetVocal into the conversations models were reading Our internal team of AI researchers discovered Reddit appearing in almost a third of ChatGPT’s search slots when searching for information. Part of our off-page strategy was informed by these searches, allowing us to target the exact conversations that were shaping the category narrative. We also used targeted link building to help GetVocal build domain authority and create the positive brand & entity associations we wanted to come through in AI answers. (4) Implemented the technical work directly We have an in-house team of engineers and data scientists which means we can find deeply technical issues and also fix them. For GetVocal specifically we created around 20 pull requests (PRs) in GitHub for their team to merge. // P.S. Want to get more customers from SEO/AEO? Hypergrowth SaaS companies like Granola, Instantly, Composio and others trust Discovered Labs as their organic search partner. We don't have account managers. We're a team of AI/ML engineers, senior SEOs and editors embedded within your team and held accountable for pipeline. 📌 If you'd like to have a quick chat with me, please go here: https://lnkd.in/eHVf23k2

AdvisoryAI's demos and closed won revenue are up 100% YoY from SEO/AEO. We've been working with them over the last several months. I caught up with Ben who runs marketing at AdvisoryAI, a pre-Series A startup building AI tools for financial advisers. One of my fav pieces of feedback from Ben: "Other agencies inside our shared Slack channel mentioned they want their communication standard to be at the same level as Discovered Labs" When we took over from their previous SEO agency, a site move had seemingly cost them nearly all their organic traffic. They ranked for very few keywords and the majority of clicks came from people already searching for AdvisoryAI by name. Here's what we changed: (1) Fixed the foundation before anything else The site had real technical issues. We ran a site audit and shipped the fixes within a day or two which almost immediately had an impact. (2) Used proprietary AI visibility measurement to build the content roadmap Everyone's using AI visibility tracker atm (or trying to vibecode one over a weekend). The problem is that many of them rely on flawed measurement methods. With our approach, we were able to get hyper-targeted on the topics we went after, helping them increase share of voice by 244% for a key topic cluster in one month. Every piece of content managed E2E by our team of editors, structured using our CITABLE framework to ensure the best chance of performing across Google and LLMs. (3) Used Reddit to influence the sources AI draws from Most people judge Reddit's value by the citations visible inside ChatGPT. We intercepted ChatGPT's retrieval traffic and found Reddit occupied roughly 27% of its search slots, despite only appearing in a small % of visible citations. So we focused on the Reddit threads already shaping the commercial prompts that mattered to AdvisoryAI, with one of our placements becoming the most-cited Reddit source LLMs are drawing from. (4) Used data to find and go after competitor gaps By July, AdvisoryAI had overtaken one of their main competitors. We were able to do this by drilling down into the specific topic clusters and model-level performance, identifying where competitors were weak, and putting together a dedicated strategy for displacing them. Ben's own words: "You guys are able to take what everybody else refers to as a black box, which is LLM search results, and make it a white box." Now, we're about to go even deeper with original thought leadership so we can create distance :) // P.S. Want to get more customers from SEO/AEO? Hypergrowth SaaS companies like Granola, Instantly, Composio and others trust Discovered Labs as their organic search partner. We don't have account managers. We're a team of AI/ML engineers, senior SEOs and editors embedded within your team and held accountable for pipeline. If you'd like to have a quick chat with me, please go here: https://lnkd.in/eHVf23k2

We helped Gladia 7x their sales-accepted leads (SALs) in 4 months. 90% of them came from LLMs, not Google. Gladia sell audio infrastructure to developers. About as niche and technical an audience as it gets, which is the worry every SaaS company has: can an agency actually understand our product and write for a technical audience? (engineers are allergic in to marketing) Here's what we did: (1) We audited the whole picture before changing anything → technical health (indexation, crawlability, speed), fixed by our in-house AI/ML engineers → on-page and content gaps, mapped against what buyers ask → benchmarked Gladia against competitors across ChatGPT, Claude, Perplexity and Google using our in-house tech We model each client's domain on their own first-party data (GSC, GA4, call transcripts) so we start from commercial priorities, not a generic best practice checklist. (2) We turned the diagnosis into a 90 day plan → mapped the pillars and clusters Gladia needed to own → sequenced it across 30, 60 and 90 days → ran it off a shared calendar with their in-house team so we didn't overlap Everything prioritised by impact. (3) We published content that competes on Google and LLMs This is where a niche audience worked in our favour: → narrow positioning means we can target the specific, long-tail prompts that map to how Gladia sells → we pulled real code snippets straight from their API docs, which gives the content natural information gain loved by Google and LLMs → we embedded real UGC from users that adds credibility and provides consensus on claims (LLMs need this to avoid hallucinations) → every piece structured on our CITABLE framework so a model can lift a clean answer and cite it → AI-assisted for scale, owned end to end by our editorial team with a tight feedback loop to maintain the high bar for content quality Information gain is your defence against every algo change coming. (4) We earned the off-page signals that shape AI answers → target Reddit threads and subreddits surfaced by our own AEO measurement → our FACTUAL framework keeps us feeding models the right signals so the company is being positioned well against competitors → grow share of voice on the subreddits Gladia's buyers read Four months in: 7x sales-accepted leads, 90% of them from LLMs. Brilliant team and we're only getting started! Right now AI search is the cheapest and easiest it will ever be. That edge has a shelf life 🙂 -- P.S. Want to get more customers from SEO/AEO? Hypergrowth SaaS companies like Granola, Instantly, Composio and others trust Discovered Labs as their organic search partner. If you'd like to have a quick chat with me, please go here: https://lnkd.in/eHVf23k2