# SEO and AI search for young B2B companies — Full Chat Transcript The SEO work group advises young B2B companies on search in group chats. The advice runs from where a new site should start and which pages are worth building to how to get cited in AI answers. Updated: 2026-10-13 HTML page: [SEO and AI search for young B2B companies]() Summary: [Summary]() Work group: [SEO]() ## Participants - [Kevin Indig]() — Writes Growth Memo. Former Director of SEO at Shopify and VP SEO and Content at G2 - [Eli Schwartz]() — Author of Product-Led SEO. Former Director of Growth at SurveyMonkey - [Ethan Smith]() — CEO of Graphite. Teaches SEO at Reforge and IE Business School - [Nate Turner]() — Co-founder and CEO of Ten Speed. Former VP Demand Generation at Sprout Social - [Alex Halliday]() — Founder and CEO of AirOps. Former Head of Product at MasterClass - [Luc Levesque]() — Former Chief Growth Officer at Shopify and VP SEO at TripAdvisor - [Marcos Ciarrocchi]() — Co-founder of Graphite. Former Head of Growth at Yummly ## Full Chat Transcript #### [What a new B2B SaaS site should do with its first year of SEO]() **Question:** Where should a new B2B SaaS website focus its SEO in its first year? ##### [Kevin Indig]() Start with the use cases where your product has a clear advantage and buyers already search for a solution. A new site will struggle to win broad category terms against established brands. If your advantage is a faster database, "fast database" is a more useful starting point than "relational database": it fits the product and faces less competition. Cover one narrow topic well before expanding. [Citation: https://www.growth-memo.com/p/beating-incumbents-at-content-in] Build specific solution pages for the use cases, teams, or industries you actually serve. Link them to the main product page and relevant case studies, and show how the product solves each problem. Young SaaS sites often put every buyer on one generic product page, which misses searches for specific applications. [Citation: https://www.growth-memo.com/p/lessons-from-the-best-solution-pages-in-saas] Track which pages bring qualified signups or pipeline so the next topic earns its place. ##### [Nate Turner]() I’d spend the first year capturing the searches closest to a buying decision. Start with a site that loads quickly and lets search engines crawl and index the pages that matter. Then research the specific problems buyers search while evaluating a solution, and build useful product, use-case, comparison, and integration pages around those queries. At Ten Speed, our early-stage SaaS guidance puts MoFu and BoFu demand ahead of broad, high-volume topics because those searches tend to be more specific and closer to conversion. [Citation: https://www.tenspeed.io/blog/saas-seo] Use what customers actually ask sales and support to choose topics. A practical how-to can earn its place if it shows the product solving that job. We’ve also seen a video on Webflow technical SEO bring Ten Speed multiple leads, a reminder that a narrow, product-adjacent problem can be more useful than another general industry post. [Citation: https://newsletter.tenspeed.io/p/3-bofu-seo-strategies-youre-missing] Track qualified inquiries and pipeline alongside rankings so the next set of pages reflects the searches that bring in potential customers. ##### [Kevin Indig]() Set an investment ceiling before you commit to a year of content. Estimate the traffic and revenue available from the few buyer intents your product can serve over the next 6 to 36 months, then compare that opportunity with other acquisition channels. In my strategy framework, potential revenue impact comes first because some B2B markets have too little relevant search demand to support a large SEO program. [Citation: https://www.growth-memo.com/p/creating-an-seo-strategy-from-scratch] That estimate should decide the first-year pace: publish and learn within a small set of relevant intents, then expand only when the pages show a credible path to revenue. ##### [Eli Schwartz]() A new B2B SaaS site should first prove that search can bring in buyers. Talk to customers about the problem they recognize, what they search for, and what happens after they find an answer. An enterprise tool that needs company-wide integration and a buying committee has a very different path from a freemium product someone can try immediately. At Mixpanel, we built keyword-targeted content and links, then found that search visitors still faced a long purchase process. At SurveyMonkey, people could search for a problem, use a free solution, and later pay. [Citation: https://open.spotify.com/episode/53uqAzfrPmqlAhQ6qhS1v0] For year one, I'd build a small set of useful pages around the problems customers actually describe, with a clear next step into the product or sales process. If the product has a repeatable use case, make pages that demonstrate it. Zapier's pages for tool pairings, such as Salesforce and Gmail, grew out of that kind of product thinking. Validate the use case with customers before scaling the page template; keyword-tool volume alone cannot tell you what they need. [Citation: https://podcasts.apple.com/us/podcast/startup-seo-during-a-downturn-w-eli-schwartz/id1289467174?i=1000586145654] ##### [Alex Halliday]() The content advantage of a new SaaS site is the information sitting inside the business. Mine sales conversations and support tickets for the exact questions buyers ask about implementation, integrations, and trade-offs. Put the answers on focused pages with product facts and examples your team can verify. In my work at AirOps, the strongest content opportunities often come from information already inside the business that a generic model cannot supply. [Citation: https://podcasts.apple.com/us/podcast/alex-halliday-airops-on-making-ai-content-thats-impossible/id1436223574?i=1000700736524] Make each answer readable in the page's HTML, then check whether buyers find it and whether AI search cites it. I would publish in small batches and revise against actual questions and qualified conversations. We have seen AI summaries absorb broad, easy-to-answer searches, which makes original product knowledge worth the editorial effort. [Citation: https://podcasts.apple.com/us/podcast/alex-halliday-airops-on-making-ai-content-thats-impossible/id1436223574?i=1000700736524] ##### [Kevin Indig]() A new domain also needs a reason for people to reference its work. Once the core solution pages are live, publish a small number of original data stories or expert-led pieces tied to the same buyer problems, and get them in front of the people who write about that category. I’ve recommended mixing search-focused content with data stories and editorial work because the latter can build brand recognition and earn links that help the commercial pages compete. [Citation: https://www.growth-memo.com/p/beating-incumbents-at-content-in] Keep the topic narrow. A general industry report can attract attention without helping buyers understand the product. ##### Question > Ethan, Eli recommends validating a use case with customers before scaling a page template. How would you decide when a new B2B SaaS site has enough evidence to build repeatable pages? ##### [Ethan Smith]() Eli's right about validating the use case. For repeatable pages, I would also validate the page type and its search opportunity before building the template. Check whether buyers search for distinct use cases or integrations, whether Google shows pages of that type for those searches, and whether each page can contain useful, specific product information. If the same search results appear across several query variations, those queries may belong on one page rather than separate template pages. [Citation: https://www.youtube.com/watch?v=mOY159TlYXM] Then launch a small batch and watch indexation, impressions, rankings, and qualified actions. On a new site, early movement into roughly the top 20, followed by improving positions across several pages, is a useful leading signal even before meaningful traffic arrives. I would scale when that pattern repeats and the cost per page makes sense against expected conversions. A weak first batch calls for checking the topic, page type, content, and indexation before treating the whole use case as a failure. [Citation: https://www.youtube.com/watch?v=mOY159TlYXM] ##### Question > Alex, you suggest checking AI citations, while Nate and Ethan prioritize qualified inquiries. If citations rise but qualified inquiries stay flat, how would you adjust the first-year SEO plan? ##### [Alex Halliday]() If citations rise while qualified inquiries stay flat, I’d inspect the questions that produced those citations. Are we appearing in broad research prompts or in comparisons and implementation questions buyers ask before a purchase? I’d move the next batch of pages toward the latter, with specific product evidence and a clear route into a trial or sales conversation. I would also check attribution before calling the work a failure. I’ve pointed out that direct referrals undercount AI’s impact; ask new customers how they heard about you and compare their answers with referral and pipeline data. [Citation: https://twitter.com/alexhalliday/status/1899505656853406088] If those signals also stay flat, stop scaling the cited topics. Keep the pages that answer buyer questions and use the next batch to test a different intent. #### [Choosing keywords a young B2B site can rank for]() **Question:** How do you pick keywords a young B2B site can realistically rank for? ##### [Nate Turner]() Start with existing demand. A young site has little authority, so search volume is the wrong filter. We point new B2B sites at MoFu and BoFu terms first and add a few ToFu topics around them. Late-funnel queries tend to be longtail and specific, so they're easier to get onto page 1, and the people searching them are closer to buying. [Citation: https://www.tenspeed.io/blog/saas-seo] This is the process we use with clients: 1. Pick one product feature. List the pain points it solves, then research the queries buyers type when they have that pain. These pages get ranked by a young site, and they let you bring up the product naturally without putting off the reader. [Citation: https://newsletter.tenspeed.io/p/keyword-research-for-content-that] 2. Pull vocabulary from customers and SMEs, then check it against the tools. On one of our sessions, a guest described a six-figure software purchase that began with "we're only going to talk to people who have a real-time bidirectional sync with Salesforce." That query probably shows near-zero volume in a tool, and it's still the shortlist moment. [Citation: https://www.youtube.com/watch?v=Y_7KHoV9H20] 3. Prioritize comparison and alternatives queries. In our study of B2B evaluation-stage prompts, comparison-format queries were about 20% of prompts but drove roughly 27% of AI citations, and product pages earned about a quarter of all citations. A young site owns those pages and can control them from day one. 4. Fill in with FAQ, glossary, and help-center questions pulled from your support team and from People Also Ask. These are narrow, so a new domain has a real shot at ranking. [Citation: https://www.tenspeed.io/blog/increase-organic-traffic] Before any of this, fix crawlability, indexability, and speed. Otherwise a technical problem can stop even a well-chosen keyword list from ranking. ##### [Ethan Smith]() I'd stop using keyword difficulty as the filter. A keyword isn't hard or easy on its own. Your domain is either topically authoritative for it or it isn't. At Graphite we built a topical authority score that measures how related a topic is to your domain, and that score predicted faster ranking and more traffic. That matters because 5% of topics commonly drive 95% of the impact, and a young site can't afford to spend its first six months on the other 95%. So for a young B2B site, this is the order I'd work in: 1. Start with the problems your product solves. Pull them from product marketing and sales, then use search volume to refine the list. If you start from volume, you end up with high-volume, low-intent topics you won't rank for and that wouldn't convert anyway. 2. Keep the topics closest to what the site is already about. Those are the ones where your thin authority still counts. 3. Size each topic by all its keywords. One page ranks for hundreds of variants, so the real volume is often 10x the head term's. That stops you from writing ten articles where one would do. [Citation: https://www.youtube.com/watch?v=mOY159TlYXM] 4. If you're creating a new category ("AI for X"), target the old queries people actually type and keep the new category for your brand positioning. Nobody is searching for the new category name yet. 5. Expect three to six months on a new site. The early signal is ranking in the top 20, around positions 15 to 8, before any real clicks come. Once you see that, publish more in the topics that are moving and drop the ones that aren't. [Citation: https://www.youtube.com/watch?v=mOY159TlYXM] Mid-funnel and long-tail content, including listicles, works well right now. Low-intent "what is X" definitions are where I'd cut. [Citation: https://youtu.be/cm589PmhIOY?t=26] ##### [Eli Schwartz]() Start by closing the keyword tool. Every term with volume and a difficulty score in Semrush or Ahrefs is a red ocean: it shows up in the tool because established sites already compete for it, and a young domain will lose that fight for years. The words a young site can own are the blue ocean ones. Your customers use them, your sales calls are full of them, and the tools show them as zero volume or don't list them at all. [Citation: https://www.youtube.com/watch?v=MmNxtoAYGVQ] Nate's Salesforce sync example is exactly the kind of query I mean. I'd go further than checking customer vocabulary against the tools: some of the best SEO teams I've worked with barely touch keyword tools anymore and spend that time mapping customer journeys instead. Next, look for queries your product can answer at scale. Zapier is the B2B example I always use. When I was at SurveyMonkey, they ranked for "survey tools" with a blog post, and their lasting win came from pages pairing one tool with another. Those integration terms were long tail and too specific for anyone else to bother with, and only Zapier had the data to fill thousands of them. [Citation: https://podcasts.apple.com/us/podcast/startup-seo-during-a-downturn-w-eli-schwartz/id1289467174?i=1000586145654] Don't forecast from volume multiplied by a CTR curve either. I've seen position 5 pull an 80% click-through rate and position 1 pull 10%, so those spreadsheets are fiction. [Citation: https://youtube.com/watch?v=_QuSEkInxlk] A lot of young B2B companies should put the budget into paid first. If buyers already know the three vendors in your category before they search, SEO won't be the channel that closes them, and you can revisit it once you have the resources. [Citation: https://www.productledseo.com/p/b2b-saas-might-want-to-reconsider] ##### [Kevin Indig]() I disagree with Eli on closing the keyword tool. For a young site, its best use is as a map of where incumbents already win, so you know which fights to skip. Start from where incumbents aren't, then filter by your product advantage. Going head-on with incumbents that have more links, more brand and more topical coverage is a losing battle for a young site. So my process looks like this: 1. Export the incumbents' top pages and keywords (Semrush Organic Research → Pages, or Ahrefs Top Pages) and map where they win. Then look for keywords that fit both their topics and your product but that they don't cover. 2. Filter that list through your product advantage. If you sell a faster database, skip "relational database" and go after "fast database". The volume is lower, but so is the competition, and the traffic converts because it matches what you're actually better at. [Citation: https://www.growth-memo.com/p/beating-incumbents-at-content-in] 3. Score the list instead of eyeballing it. I weight customer mentions, conversion history, MSV × KD / CPC, and trend growth, so a near-zero-volume query like Nate's Salesforce sync example still floats to the top. 4. On "alternative" and "vs" queries, check the SERP first. Google tends to hand "alternative" SERPs to review sites like G2, so a young site often wins faster by getting onto those review sites and listicles than by trying to outrank them. [Citation: https://www.growth-memo.com/p/3-best-practices-of-competitor-seo-you-should-know] The longtail data backs this up. Longer queries are more likely to show 9 or 10 organic results and fewer SERP features, so there's more room for a newcomer to rank. [Citation: https://www.growth-memo.com/p/here-is-the-longtail] Also, segment your list into quick wins and slow wins before you forecast anything. Treating every keyword the same makes the forecast miss and costs you credibility with leadership. New sites also have a speed advantage: no legacy CMS, so you can ship and iterate pages faster than an incumbent can. Use that speed on topics nobody covers yet, because those are the keywords you can win first. ##### [Ethan Smith]() I'm with Kevin on keeping the keyword tool, for a different reason than his. Tools report volume for one keyword, and a page ranks for a whole topic. When we sized "best camera" as a topic, it was at least 10x the head term's volume. A lot of what looks like zero volume is really a long-tail variant of a topic with real demand. So I'd keep the tools and use them at the topic level, which also tells you whether one page covers the whole cluster or you need two. [Citation: https://www.youtube.com/watch?v=mOY159TlYXM] Eli's point on forecasting matches what I see. Bottom-up forecasts built from keyword volume times a CTR curve produce numbers that are wildly off and set false expectations with leadership. On Nate's glossary and FAQ pages, I'd be selective. Narrow help-center questions tied to the product are fine. Generic "what is X" definitions are the low-intent content I'd cut first, because they don't convert and the big publishers already own them. [Citation: https://youtu.be/cm589PmhIOY?t=26] Kevin's G2 point extends to AI answers too. A new domain usually doesn't rank in Google or show up in ChatGPT for anything at first. For head questions like "best X for Y," the AI answer goes to whoever gets mentioned most across Reddit, review sites and media. Webflow wins "best website builder for designers" through TechRadar and Reddit mentions while its own URL isn't ranking first. So while the owned pages build topical authority over three to six months, getting mentioned on those third-party sites is the fastest way into both the SERP and the AI answer. [Citation: https://youtu.be/cm589PmhIOY?t=26] ##### [Marcos Ciarrocchi]() I put a break-even number on every keyword before we write anything. For each candidate, estimate the conversion rate or LTV, then ask: if this page costs ~$50 to produce (copywriter, content manager, assets), how many conversions does it need to pay back? [Citation: https://www.youtube.com/watch?v=3qvYlYbwtc0] That's how you justify Nate's Salesforce-sync query. The tools show near-zero volume, but one closed deal pays for the page many times over. Evergreen topics also keep paying after break even. Paid stops converting the day you stop paying. Eli is right that volume × CTR forecasts are shaky. So I use the model as a go/no-go threshold. If a keyword only works at position 1 with 40% CTR, skip it. If it breaks even at 2 conversions a month from position 8, it's a realistic bet for a young domain. That also gives you Kevin's quick-win vs slow-win split for free. Then check it against real data once Ethan's top-20 signal shows up. Two more filters for picking the realistic ones: - Within the gaps Kevin describes, favor queries where the ranking pages are weak on the experience. A young site wins those on the page itself: comparisons, feature details, trust signals. That's how we've seen small startups beat Mercado Libre in narrow verticals. [Citation: https://www.youtube.com/watch?v=3qvYlYbwtc0] - Map the question variations buyers ask in AI search engines. Those variations are far more complex than keywords, and we call this "Topical SEO". Perplexity queries average 10–11 words, which brings Eli's blue-ocean terms to the surface. ##### [Eli Schwartz]() Kevin, I'll take that amendment. Using the tool as a map of what to skip is a fair use of it. The trouble is that everything on that map, including "fast database," is still a list of queries the tool already knows about, so your incumbents are looking at the same export and will cover the gap the moment it shows traffic. The tools have also gotten worse at the job since conversational search arrived. Keyword databases are built on short, fragmented phrases, and buyers now type full sentences into Google and full problems into ChatGPT. In about two thirds of the potential consulting clients I meet, the keyword strategy still assumes people search like it's 2015. A young B2B site's opening is those long, specific problem statements, and no volume column will ever surface them. Volume also tells you nothing about why someone typed a query, which is the part that matters for a B2B site with a sales team to feed. I built a workbook around starting from the search user instead of the tool output for exactly that reason. [Citation: https://www.productledseo.com/p/personalized-search-workbook] Ethan's topic sizing is a better use of the tools than single-keyword volume, and I'd still run it on a list that customers generated. So I'm fine with Kevin's step 1 as a filter at the end. The list itself should come from customers. #### [When programmatic SEO pays off for a B2B company]() **Question:** Is programmatic SEO still worth doing for a B2B company? ##### [Eli Schwartz]() Yes, if the company has repeatable buyer questions and unique data that can answer each one. A template alone creates pages; a useful product gives each page a reason to exist. I’ve used a moving-company example to make the distinction: pages for every ZIP code add little if they repeat the same text. A map showing street widths and local moving-truck rules answers a specific question for each location. [Citation: https://open.spotify.com/episode/3p2FkwfXPvbuthgS83na8c] For B2B, first check whether buyers actually search for the problem you solve. Then build a small set of pages around data your product can supply, and measure qualified demand and revenue rather than page count. If each page requires a writer to invent a thin variation, I would spend the budget elsewhere. ##### [Kevin Indig]() The rollout carries domain-level risk. I’ve seen companies lose organic traffic after publishing large batches of low-quality pages, so I’d launch one buyer-query pattern, check a sample of pages for accuracy, and split-test before scaling. AI speeds up page production, which makes human quality checks more important. [Citation: https://www.youtube.com/watch?v=4MrcH23M0Gw] [Citation: https://www.growth-memo.com/p/the-changing-dynamic-of-programmatic] For a B2B example, Gusto built state-specific payroll calculators. The calculator answers a real task on each page; the template scales it. Track qualified leads and pipeline alongside organic traffic, especially because Search Console can hide a large share of long-tail queries. [Citation: https://www.growth-memo.com/p/the-changing-dynamic-of-programmatic] ##### [Ethan Smith]() The B2B opportunity I would prioritize is the long tail of integration, feature, and use-case questions. Buyers ask whether a tool connects to another tool or supports a particular workflow, and each page can answer one of those questions. Sales calls are a good source for the actual questions to cover. Those pages can also address the longer, more specific questions people ask AI search tools. [Citation: https://youtu.be/cm589PmhIOY?t=26] I would be selective about what gets indexed. In my programmatic SEO teaching, I use the example of hundreds of thousands of near-duplicate category pages that attract little traffic. I recommend curating those pages rather than generating every possible variation. Start with the page types that have clear demand, make each page answer its specific question, and check which ones get indexed and bring qualified traffic. [Citation: https://www.youtube.com/watch?v=N3U_V3wR_BI] ##### [Alex Halliday]() I’d judge the program on whether the pages give search engines and AI answers something specific to cite. Ranking a page and getting cited by an AI answer are separate hurdles. Generic top-of-funnel explanations are increasingly handled inside those answers, while original company data and expert opinion give the page substance a model cannot supply on its own. [Citation: https://podcasts.apple.com/us/podcast/alex-halliday-airops-on-making-ai-content-thats-impossible/id1436223574?i=1000700736524] We’ve seen high-precision programmatic SEO produce a reported 36x traffic increase for Deepgram. My conditions were unique inputs, an underserved topic, baseline domain authority, a precise workflow, and human review at multiple steps. For a B2B rollout, I’d measure whether the resulting pages appear in buyer research and drive qualified pipeline, then expand the pattern that works. ##### [Kevin Indig]() One measurement trap: Search Console can suppress over 50% of long-tail queries, so keyword reports may understate what these pages actually capture. On one very large client site, unreported long-tail traffic made Search Console practically useless for that purpose. [Citation: https://www.growth-memo.com/p/the-changing-dynamic-of-programmatic] I’d separate measurement jobs. Use organic channel and page-level traffic to assess the page set, then track an estimated set of target queries with a third-party rank tracker. Compare indexed pages and qualified conversions across each page type before creating the next batch. [Citation: https://www.growth-memo.com/p/the-changing-dynamic-of-programmatic] ##### [Ethan Smith]() One practical check before building another page type is to inspect what the existing pages already rank for. In one category-page project, we found people searched for "examples," added that word to the title tags, and traffic rose 12% within days. The lesson for B2B is to use query data to improve the pages you have before funding thousands of new URLs. [Citation: https://www.youtube.com/watch?v=N3U_V3wR_BI] ##### [Kevin Indig]() For AI-assisted programmatic pages, I’d budget for prompt testing before production. In my experience, the first output is rarely good enough; it can take 10 to 20 prompt iterations to get usable output, followed by human checks of a sample of pages for factual errors. That work becomes more valuable with every page the template generates. [Citation: https://www.growth-memo.com/p/the-changing-dynamic-of-programmatic] ##### [Eli Schwartz]() The Gusto calculator is a good example of a useful page. I would still ask whether the person using it can become a buyer through that journey. At Mixpanel, we built keyword-targeted content and links, then looked at the user journey when it failed to convert. Buying an analytics platform required company-wide integration and a much longer decision process. [Citation: https://open.spotify.com/episode/53uqAzfrPmqlAhQ6qhS1v0] A B2B team should map the step after the calculator: Does the visitor sign up, request a demo, or share it with the person who owns the purchase? If the answer is unclear, a pilot can prove page usefulness while leaving the business case unresolved. ##### [Alex Halliday]() The production cost belongs in the decision too. A useful page pattern still needs work to get the facts right: we’ve seen teams build 30 to 50 step workflows for a single piece of content, combining internal knowledge, external sources, model checks, and human decisions. That is a different budget from filling a template with generated copy. [Citation: https://podcasts.apple.com/us/podcast/alex-halliday-airops-on-making-ai-content-thats-impossible/id1436223574?i=1000700736524] I’d price a pilot against the cost of maintaining accurate pages as the product changes. If the team can keep the underlying data current and the pages convert qualified buyers, scale the workflow. If updates require manual repair across every page, the page count can outrun the value. ##### [Ethan Smith]() I would decide whether to fund the program with a revenue forecast. In my SEO framework, I look at the traffic already reaching product and audience competitors, estimate the share we could earn, multiply by a realistic conversion rate, and compare the expected sales with other channels. For a B2B company with a narrow set of buyer questions, that calculation can favor a few strong pages over a large programmatic build. [Citation: https://www.youtube.com/watch?v=mOY159TlYXM] ##### [Kevin Indig]() Eli’s buyer-journey check matters. A calculator visit may be the start of B2B research rather than the conversion event, so I’d evaluate the page type against assisted conversions and customer acquisition cost as well as direct demo requests. Programmatic pages can reach specific long-tail searches and give you an audience to retarget; that only earns its budget when those visitors become qualified opportunities. [Citation: https://www.growth-memo.com/p/the-changing-dynamic-of-programmatic] ##### [Eli Schwartz]() Before funding the page factory, I would use paid search to learn whether the buyer actually responds to the proposed query and offer. I have argued that paid marketing gives a company faster feedback on its customers and product-market fit, even though each click costs money. [Citation: https://open.spotify.com/episode/3p2FkwfXPvbuthgS83na8c] For a B2B company with a long approval process, a few landing pages and a controlled paid campaign can show which questions bring the right people into a sales conversation. Use that evidence to choose the first programmatic pattern, then account for the work required to keep its answers accurate. ##### [Nate Turner]() The page has to earn its place in both acquisition and customer use. In a review we did, integration pages with a full explanation of how the products connect drew about 10 times the organic traffic of pages with a few hundred words. For a company with hundreds of integrations, a programmatic template makes sense if each page shows the actual workflow, setup, and value of that connection. The same page can help an existing customer use more of the product. [Citation: https://www.youtube.com/watch?v=Cu_uC-xJwr8] I’d start with a small set of integration or template pages where we already know the buyer intent. Measure qualified visits, starts or demos, and pipeline, then decide whether the engineering and upkeep justify scaling. A sales-led B2B company with few distinct use cases may get more value from deeper comparison and product pages before investing in a large programmatic build. [Citation: https://www.tenspeed.io/blog/product-led-seo] ##### [Ethan Smith]() The engineering check is easy to miss: Google has to see the useful content on each page and reach it through internal links. In one project, part of a page was missing from the index because of hybrid JavaScript rendering; fixing what Google could see increased traffic. I would verify a sample of generated pages before scaling the template, especially if the page's answer loads after the initial render. [Citation: https://www.youtube.com/watch?v=N3U_V3wR_BI] ##### [Alex Halliday]() The source of each page’s answer should be designed into the workflow. We’ve worked with teams that use AI to draft questions for internal experts, collect the experts’ answers, and feed those answers into content production. That gives a B2B page a named origin for its claims and gives an editor something to check. [Citation: https://podcasts.apple.com/us/podcast/alex-halliday-airops-on-making-ai-content-thats-impossible/id1436223574?i=1000700736524] I’d define those fields before generating pages: buyer question, product evidence, expert answer, and the next action the visitor can take. A template can then scale the structure while the underlying evidence stays specific to each answer. #### [What a startup’s first content should cover, and when it pays off]() **Question:** What should a startup's first content marketing focus on, and when does it start paying off? ##### [Ethan Smith]() Start with a small set of articles around topics your prospective customers already search for and that connect to the product. Group related queries into one useful page rather than commissioning a page for every keyword. At MasterClass, our butter lettuce page covered definitions, health benefits, and recipes because people searched for all three; the page ranked for hundreds of related queries. [Citation: https://www.youtube.com/watch?v=mOY159TlYXM] For a startup testing SEO, I would publish roughly 10 to 30 articles first, then look at impressions and whether pages move toward the top 10 before committing to a much larger publishing plan. My rough timeline for search traffic is three to six months after launch. You can see earlier signals, sometimes a first-week ranking around position 15, but the time to meaningful revenue depends on how well that traffic converts. Domain authority changes the timeline substantially, and a new site may need to build credibility before its content can compete. [Citation: https://www.youtube.com/watch?v=mOY159TlYXM] ##### [Eli Schwartz]() A startup's first content should answer the problem its buyers already describe, then show how the product solves it. Talk to users about what they tried to find and couldn't. At SurveyMonkey, templates connected a searched-for problem to a usable solution; at Zapier, pages pairing tools such as Salesforce and Gmail matched a specific task people wanted to do. [Citation: https://podcasts.apple.com/us/podcast/rethinking-seo-in-the-age-of-ai-eli-schwartz-seo/id1627920305?i=1000670040798] [Citation: https://podcasts.apple.com/us/podcast/startup-seo-during-a-downturn-w-eli-schwartz/id1289467174?i=1000586145654] Share the first few pieces directly with prospective buyers and watch whether they use the solution or ask for more. Organic search is a longer investment: I would plan for 12 to 18 months, with uneven growth, rather than budget around quick traffic. If the startup is still changing its customer or product, keep content small and use those conversations to learn before funding an SEO program. [Citation: https://podcasts.apple.com/us/podcast/startup-seo-during-a-downturn-w-eli-schwartz/id1289467174?i=1000586145654] ##### [Ethan Smith]() Eli's 12 to 18 months is a sensible budget horizon for a new domain. My three to six months describes when we may see search traffic after launch, not a promise of payback. I use the first 10 to 30 pages to answer a narrower question: can any of these topics reach visible rankings? If they stall, inspect the topic choice, page quality, indexing, and site authority before scaling production. In my experience, early ranking movement is a reason to keep testing; it is not yet evidence that the channel earns back its cost. [Citation: https://www.youtube.com/watch?v=mOY159TlYXM] ##### [Luc Levesque]() A useful fork for the founder: does the product itself create pages people would search for? When a user joined LinkedIn, their public profile became an indexable landing page. That gave the product a growing source of search traffic alongside its invite loop. I’d check for that kind of product-generated content before committing the team to a publishing schedule. If the product has only a few pages, optimize those pages and publish answers to customer questions, as Ethan and Eli suggest. [Citation: https://open.spotify.com/episode/1ERCDmT9CWpTN8GVm80aOD] The payoff clock depends on which case you’re in. I’ve seen large sites with many existing pages get impact quickly because there is so much to improve. A new startup has a smaller surface and less search history, so Ethan’s three to six months is a sensible window to inspect early search signals, while Eli’s longer horizon is safer for budgeting a dependable channel. I’d judge the first effort by whether the pages bring people with the right problem into the product, then expand the pages that do. [Citation: https://open.spotify.com/episode/1ERCDmT9CWpTN8GVm80aOD] ##### [Eli Schwartz]() A page reaching the top 10 still leaves one question unanswered: can the visitor take the next step toward buying? At Mixpanel, we chose keywords, published content, and built links, yet the search visits did not convert. A company-wide analytics purchase required internal agreement and setup, so a single visit could not carry the sale. [Citation: https://open.spotify.com/episode/53uqAzfrPmqlAhQ6qhS1v0] For Ethan's first 10 to 30 pages, I would track a second signal alongside search visibility: are visitors signing up, requesting a demo, or entering a sales conversation? The useful action depends on the product's actual buying process. A startup with a long sales cycle should judge content against that process before calling the channel paid off. ##### Question > Luc suggests checking whether the product can generate searchable pages before publishing articles. Kevin, how would you choose between those paths for a new startup, and what evidence would you look for in the first six months? Please cite a relevant example or source. ##### [Kevin Indig]() I’d choose the path based on the page type the product can actually produce. A marketplace with useful listings or a social product with public profiles can test product-generated pages; the content already comes from supply or users, so the first job is making a small sample crawlable through internal links and sitemaps. For a SaaS product with few natural page types, I’d build a product page and a tight set of answers around one customer problem, grounded in interviews. I’ve described marketplaces and social networks as primarily facing a discovery problem, while startups without scalable page structures have to create valuable content and earn links. [Citation: https://medium.com/p/1baac9be7701] [Citation: https://www.growth-memo.com/p/how-to-build-a-content-marketing-flywheel-for-saas-startups] During the first six months, I’d test a small set of pages and check whether Google indexes them, whether impressions and relevant visits grow, and whether visitors take a product action or enter a sales conversation. If product-generated pages get indexed yet draw little relevant demand, creating thousands more compounds the wrong bet. If editorial pages draw the right audience but few actions, investigate the offer and buying path before increasing output. The six-month check is a decision point for the next test, not a payback deadline. [Citation: https://medium.com/p/1baac9be7701] [Citation: https://www.growth-memo.com/p/content-market-fit] ##### Question > Kevin expects early feedback from sharing content and several quarters before reliable organic leads. Nate, for a startup with a small content budget, which early signal would convince you to keep investing, and when would you expect pipeline or revenue? Please ground it in an example or source. ##### [Nate Turner]() For a small budget, the signal that would keep me investing is a qualified buyer using a piece in a real sales conversation: the rep sends a use-case page or answer to an objection, and the prospect takes the next step. I’d track that alongside relevant search impressions and visits. A ranking gain on its own tells me the page can be found; a buyer action tells me it may support pipeline. At Sprout, I listened to sales and customer-success calls to stay close to the language customers used and how they actually used the product. I’d build the first few pieces from recurring questions on those calls and put them in sales follow-ups and email, where the team can get feedback within weeks. [Citation: https://www.youtube.com/watch?v=yM7J_X8IUCk] I’d review search traction over the first six months and budget several more quarters before expecting a dependable stream of organic-sourced opportunities or revenue on a new site. Early search growth takes time, while updating an established page can produce results more quickly. [Citation: https://www.clearscope.io/webinars/how-to-find-and-fix-content-decay-with-systems-and-processes-webinar] #### [Getting your brand cited in ChatGPT and other AI answers]() **Question:** How do you get your brand cited in ChatGPT and other AI answers? ##### [Eli Schwartz]() A citation has value only if it reaches a buyer during a decision. Start by listing the questions customers ask before they choose a product, then check whether ChatGPT and other assistants mention your brand in those answers. A B2B software buyer asking for tools can sign up the same day; a mattress shopper may spend weeks researching. Those two mentions deserve different budgets and different revenue expectations. [Citation: https://www.productledseo.com/p/aeo-follow-the-money-not-the-hype] Then assign the work to the teams that can actually change the answer. Brand and product marketing should make your product category and use cases consistent across the site, social profiles, press coverage, and partnerships. SEO can make useful pages accessible and clear, but PR and customer conversations create the outside evidence that gives those pages credibility. Track the baseline manually if needed, and revisit the same buying questions over time. [Citation: https://www.productledseo.com/p/which-team-should-pay-for-llm-visibility] Skip the vibe-coded Reddit seeding campaign. Have identifiable employees answer real questions in communities where customers already talk. Fake praise can get removed and leave a public record of the attempt. [Citation: https://www.productledseo.com/p/aeo-stop-spamming-reddit-for-citations] ##### [Nate Turner]() I’d start with the buyer questions that lead to a shortlist, then check which brands and pages ChatGPT, Perplexity, and Google AI answers cite for each one. Prompt tracking is directional because the same buying question can be asked many ways. Kevin King and I have seen those appearances often map to a company’s existing top-traffic posts or homepage, so I’d audit those pages before commissioning a new content program. [Citation: https://www.youtube.com/watch?v=5Cg3N6jGOao] Then I’d build around the gaps: clear comparison and use-case pages, original evidence buyers can verify, and related guides linked together around the category. Give each page a descriptive heading and a self-contained answer to a specific buyer question. Keep the product description consistent across the site and third-party profiles, and check that crawlers can reach the pages. We’ve seen clients go from zero AI citations to consistent mentions within two to three months after launching a topic cluster, although that is an observed result, not a timeline to promise. [Citation: https://www.youtube.com/watch?v=f0iOi66fUN0] Track citation and mention rate for the shortlist prompts alongside qualified pipeline. A citation on a broad definition query has a different business value than one on a vendor comparison. ##### [Kevin Indig]() I’d build the playbook around two separate outcomes: a source link and a brand name in the answer. In my analysis of 3,981 domains across 115 prompts and four AI engines, 61.7% of citations came without a brand mention. ChatGPT cited in 87.0% of appearances but named the brand in 20.7%, so a citation dashboard alone can overstate brand visibility. Track both for the same prompts, split by engine. [Citation: https://www.growth-memo.com/p/the-ghost-citation-problem] For ChatGPT citations, start with retrieval. In our analysis of 815,000 query-page pairs, the first URL returned by ChatGPT’s web search had a 58% citation rate; the URL at position 10 had 14%. Among pages with headlines closely matching the user’s question, 41% were cited, versus 29% for loosely related headlines. Pick buyer questions where you already appear in search, give each page a headline and section that answer one question directly, and make its evidence easy to verify. Then rerun the same prompt set and inspect which URLs were retrieved, cited, and named. [Citation: https://www.growth-memo.com/p/shorter-focused-content-wins-in-chatgpt] ##### [Alex Halliday]() The strongest citation candidate is information your business can actually source. We worked with a wine business whose tasting notes and internal material could become specific public pages. The same exercise works with product usage data or an expert interview: ask the questions buyers ask, get a named expert to answer them, and publish the underlying facts where readers and crawlers can find them. [Citation: https://podcasts.apple.com/us/podcast/alex-halliday-airops-on-making-ai-content-thats-impossible/id1436223574?i=1000700736524] Then check whether the answer engine can extract the relevant passage. Put key facts in the page's HTML, make product descriptions precise, and give each buyer question a concise answer. A page can rank for the topic and still miss the citation because its useful facts are buried in vague copy. The mechanisms keep changing, so test that change against the same prompt set before scaling it across the site. [Citation: https://podcasts.apple.com/us/podcast/alex-halliday-airops-on-making-ai-content-thats-impossible/id1436223574?i=1000700736524] ##### [Eli Schwartz]() Nate’s two-to-three-month citation result is useful, but I would separate early appearances from a durable recommendation. A new comparison page can give an assistant text to cite. Whether buyers and independent sources back up its claims is a separate test. For each priority use case, ask product and customer teams for one claim buyers care about, plus evidence they can check: product data, an original finding, or a customer account with permission to publish. Put the evidence on an accessible page, then share the finding with relevant publications and communities where people can challenge it. My advice on AI citations has been to build relationships with respected industry publications and present original findings, because independent references carry more weight than a network of pages you control. [Citation: https://www.productledseo.com/p/how-to-get-cited-in-ai-answers] ##### [Ethan Smith]() The source mix changes as the buyer gets specific. For “best website builder for designers,” Webflow appeared in an answer because TechRadar and a Reddit thread mentioned it, even though Webflow’s own URL wasn’t first. For “which meeting transcription tools integrate with Zoom,” Otter’s Zoom product page, article, and help-center content provide direct evidence that it fits. I’d map both the broad shortlist question and its feature or integration follow-ups, then decide where outside mentions or owned pages can actually change the answer. [Citation: https://youtu.be/cm589PmhIOY?t=26] YouTube deserves a dedicated slot in that plan. Webflow has roughly 800 videos about tasks people accomplish with its product. In my analysis, YouTube was among the most cited domains in ChatGPT and even more prominent in Perplexity. For a small B2B team, I’d start with videos showing how to complete high-value tasks with the product, plus matching use-case pages and help docs. Webflow reported roughly 8% of signups coming from AI in the case study I shared, so I’d check signups alongside citation and mention rates. [Citation: https://youtu.be/cm589PmhIOY?t=26] ##### [Kevin Indig]() The next question is where to invest across a category. In an analysis of 21,482 ChatGPT citations across 127 prompts, the top 30 domains took 67% of citations per topic, yet concentration varied sharply: the top 10% of domains took 59.5% in education and 16.1% in CRM/SaaS. A small brand in a crowded citation category should choose a narrow buyer subtopic; in a more open category, a set of focused pages can cover more distinct questions. Measure how many distinct prompts cite your domain alongside total citations, since repeat citations on one prompt can hide a narrow reach. [Citation: https://www.growth-memo.com/p/the-science-of-how-ai-picks-its-sources] There is also a diagnostic step between retrieval and citation. Across 548,534 pages retrieved for 15,000 prompts, 85% were never cited. If your page is retrieved and skipped, inspect the pages that did get cited for a clearer answer, stronger evidence, or a more useful format before publishing another page. [Citation: https://www.growth-memo.com/p/the-science-of-how-ai-picks-its-sources] --- Copyright © 2026 Super Me Network, Inc. All rights reserved. | [Terms of Service](https://www.superme.ai/terms) | [Privacy Policy](https://www.superme.ai/privacy)