Venice Drop In #1: 12 Lessons on SEO, AI, and LLM Visibility with Kevin Indig

Few people have ridden the organic-growth roller-coaster quite like Kevin Indig. Over the past decade he’s steered SEO programs at Shopify, Atlassian, and G2, while also advising breakout brands such as Hims & Hers, Ramp, Dropbox, Snap, and Reddit. 

What makes his perspective rare is the range: ten years in the operator seat at high-growth companies, thousands of hours experimenting with every organic lever (search, content, community), and an instinct for knitting those channels back into product strategy and board-level KPIs.

In our first Venice Drop-In, he breaks down exactly how large-language models are rewriting the SEO playbook, why clicks will drop but intent will grow, and what founders should do this quarter to stay visible when the bots do the browsing.

Five Fast Takeaways

1. Know your customers, for real. Schedule one live convo a week, hit record, run it through an AI transcript, and highlight the exact words they use. That language becomes your keyword list, your copy, your product roadmap, everything.
2. Create hyper-specific content. Forget “best pillow.” Ship a page titled “The best pillow for warm sleepers” — or even more specific. Chase less competition and laser-level relevance.
3. Expand your web-footprint. Chase the mention, not the link. A glowing shout-out in the New York Times or a trusted YouTuber beats ten DA-90 backlinks. Track sentiment and authority; let volume take care of itself.
4. Kill the channel silos. Put SEO, paid, product marketing, and social in one growth pod, share the same KPIs, and move as a pack. LLMs grade the whole brand experience, so your org chart should too.
5. Trade click counts for buying intent. AI Overviews steal the tire-kickers. What’s left are shoppers who already know what they want. Optimize the pages that close deals, then watch conversion rates spike even as raw traffic falls.

P.S. Want more from Kevin? Check out his Growth Memo newsletter. 

The Full Drop In Q&A

Scroll on for Kevin’s answers to every question from the session, lightly edited for clarity. 

Q1: How is AI changing search and SEO?

Kevin: By rolling out AI overviews, Google has the biggest consumer-facing AI surface on the planet. Every algorithm update now folds in large-language-model reasoning, which means ranking factors evolve at LLM speed, not once-a-year white-paper speed. 

‍

As AI Overviews, SGE, and direct answers expand, publishers and brands see a familiar pattern: search impressions rise, but click-through declines. Discovery happens in the AI overviews; only high-intent users bother to click. Expect fewer visits, but far warmer leads.

In the old model, you won with sheer search engine rankings real estate. In the new model, you’ll win with sharper, purchase-ready traffic. So SEO isn’t dying, it’s narrowing. Traffic that reaches you will be smaller in quantity, larger in intent, and harder to earn unless you provide substance AI can’t generate on its own.

Q2: Between traditional SEO and visibility in large-language models, do we need totally different tactics or is it just ‘do good SEO’?

Kevin: Right now I see two camps. Camp One says, “Just do good SEO, nothing has really changed. LLMs are just another smarter channel, and everything we already do for SEO will spill over.”

Camp Two says, “User behavior is changing dramatically. LLMs understand intent much better than search engines ever did, so we need new tactics.”

Both camps are correct. Almost everything we do for SEO also boosts visibility in LLMs, and there are extra things we’re not yet doing in SEO that matter a lot for AI results.

One of the biggest levers is hyper-specialized content. The average Google query is four words; the average ChatGPT prompt is thirty. People have realized they get better answers when they ask in detail.

So you have rising demand for hyper-specific answers and products, but most companies still optimize for short-tail queries. The shift we must make is to create content tuned to very specific attributes, low-competition, and laser-focused on what a narrow segment needs.

Imagine Google queries like “best CRM” or “best CRM for small business.” Now imagine an LLM prompt: “I’m looking for the ideal CRM for my seven-person law firm in Minnesota that serves healthcare clients.” Almost nobody has content for that. That’s the opportunity.

I also believe the high overlap between SEO and LLM visibility is a transition period. Today Bing and Google ground their AI answers in search results to avoid hallucinations, but LLMs get huge volumes of direct feedback from prompts. Over time they’ll develop their own criteria. If we don’t evolve with them, we’ll get stuck optimizing for a past that no longer matters.

Age in SEO is no advantage anymore. In fact, it can be a disadvantage because you’re stuck in old patterns. Old habits slow adaptation. Newcomers can skip straight to what works now.

Q3: Are keywords, metadata, alt-text, and backlinks still relevant in 2025? 

Kevin: Keywords and basic on-page hygiene still matter, but backlinks have evolved into what I call web-footprint. It’s less about a hard link and more about being mentioned in high-authority contexts with positive sentiment.

Would you rather have a link from Wikipedia or a favorable mention in The New York Times? I’d pick NYT every time. The mention reaches more people and influences how LLMs evaluate your brand.

Tactically, measure brand mentions across the web with social-listening or PR tools. Track whether mentions grow over time and whether sentiment is positive. That’s more important than raw link volume now.

Q4: Has SEO converged with digital PR and social search? Is it now one broader discipline of brand awareness?

Kevin: Absolutely. SEO outputs have always been recommendations for other teams, but LLMs 10× the need for cross-functional collaboration. Models factor in how many people know your brand, how they feel about it, whether they leave good reviews, how customer support treats them, everything.

Startups should kill channel silos. Build growth task-forces where paid search, product marketing, social, and SEO all report to one manager and share KPIs. The old structure is outdated.

Q5: How do you drive meaningful SEO growth for a low-volume, high-consideration consumer product with little branded or generic search?

Kevin: Have at least one customer conversation per week. Record it, transcribe it, and use AI to analyze patterns: the problems they’re solving, alternatives they consider, which subreddits, YouTube channels, or magazines they read, who they trust.

SEO might mean co-creating a study with a niche publisher, or engaging deeply in a subreddit if that’s where the audience validates decisions. Low volume can be an advantage because people describe their needs very precisely, and LLMs can match that to you, if the content exists.

Q6: Search engines value time-on-site. How does engagement translate into LLM rankings?

Kevin: LLMs capture multiple engagement signals:

1. Whether a user stops prompting after the answer (good).
2. Thumbs-up or thumbs-down feedback.
3. Whether they click a cited source and then stop prompting.

Monitor traffic from LLMs in your analytics. Do those visitors bounce or convert? Improve that journey, the better their outcome (spending time on your site, not returning to the AI tool to follow up with another prompt), the more often the model will recommend you next time.

Q7: Do ChatGPT, Claude, Gemini, Perplexity, etc. operate the same way when choosing sources?

Kevin: No. A Profound.ai report showed big differences: Perplexity leans heavily on Reddit, ChatGPT on Wikipedia, and Gemini sits in between. But because each model is only months ahead of the next, LLM answers will diverge more over time.

Watch your referral analytics. If Gemini suddenly sends more traffic, study where it sources data and strengthen your presence there. Stay nimble as market share shifts.

Q8: If you start SEO with a brand for the first time, where do you allocate time and budget first?

Kevin: Start with first principles: Which pages already drive business impact? Pull a 28- or 90-day report from Google Analytics or whatever tool. List pages by conversions, pipeline, or revenue. Strip out obvious outliers like the pricing page, then analyze the remainder:

• Is it a cluster of blog posts on one topic?
• Is it a handful of landing pages?

Double-down on what’s already working before chasing new keywords. Can you improve those pages’ copy, design, or internal links? Often you can grow faster by optimizing proven assets than by creating new ones.

Q9. For founders juggling 100 things, what SEO checkpoints really matter?

Kevin: SEO is a great lever but also a huge distraction if you jump in before product-market fit. You can’t prop up a weak offer anymore, the hack-your-way-to-traffic era is over. So start with signals that you truly have something people want:

1. Brand-search momentum. Install Google Search Console (just drop in the HTML snippet). Are people already typing your brand name or brand + problem into Google? If not, fix the product first.
2. “Right pages for the right questions.” Open that Search Console report and scan the brand queries: “Brand pricing,” “Does Brand integrate with X,” “Brand vs Competitor.” If users ask but you have no page for that topic publish it. Don’t make prospects hunt on Reddit or elsewhere.
3. Only after brand basics, pick one pillar topic your company should dominate. Atlassian has Agile/DevOps, G2 owns software-review content, Shopify owns eCommerce. Define yours and build a coherent cluster before spraying keywords everywhere.
4. Expand to non-brand search once conviction is high. Map the long-tail needs inside that pillar and create the hyper-specific pages we discussed earlier.

Q10. How will Shopify’s integration with OpenAI affect product discovery?

Kevin: Shopify wants to be the product data layer across every channel, similar to Google’s Shopping Graph. Partnering with OpenAI makes sense: Shopify isn’t an aggregator like Amazon; it connects individual merchants to channels.

OpenAI needs eCommerce data, but I doubt the partnership will stay exclusive. I expect OpenAI to add Amazon, Etsy, Walmart, eBay, etc. Still, being an early data provider could give Shopify merchants a head start in AI commerce surfaces.

Q11. How can eCommerce brands use reviews or UGC to strengthen an LLM’s view of product quality?

Kevin: Start with on-site reviews, but watch how they’re rendered. Most LLMs can’t execute client-side JavaScript — Gemini is the exception —so if your review widget loads in JS, the model never sees it. Pick a provider that outputs server-side HTML, or at least progressive-enhancement markup LLMs can crawl.

Next, be cautious with aggregators like Trustpilot. They’re useful, but for some products, especially utilitarian ones where people only bother to post when something breaks, you’ll skew negative. Don’t obsess over those scores if the feedback isn’t representative.

A more leveraged play is third-party credibility:

1. Affiliates with strong YouTube channels. Video walk-throughs signal “real hands-on use” and LLMs already index YouTube heavily.
2. Publisher reviews. Sometimes affiliates and publishers overlap, but treat them separately: top-tier media outlets carry disproportionate authority in AI rankings. Make it stupid-simple for journalists to test your product—send samples, data sheets, ready-to-use assets.
3. Act on feedback publicly. When reviewers point out issues, fix them and circle back. That closes the loop and encourages follow-up coverage, which LLMs interpret as positive sentiment momentum.

Q12. With limited resources, which paid channel besides Meta, Google, and TikTok looks most promising?

Kevin: It’s tough to escape paid channels entirely; performance marketing works and, if anything, AI is making it more efficient. The one under-priced, intent-driven network I keep recommending is Reddit Ads:

• Low CPMs right now. Reddit’s ad stack is still young; they need revenue post-IPO, so bids are friendly and support generous.
• Intent > behavior. Like Google, you target users by the topics they’re actively reading, not by look-alike demographics. That tends to convert better and gives you free insight into what topics matter to your market.
• Built-in topic intelligence. Reddit just launched a trend-insights tool that shows which sub-topics are gaining traction—gold for content and product research.
• First-mover edge. Because it’s early, many niches are wide-open. Consumer brands do especially well.
• Try a paid-as-research sprint. Before going all in, run Reddit ads for 2–3 weeks at a slightly higher budget. Measure: Are we driving customers? Are sessions engaged? If the signals are good, throttle spend to a sustainable level and spin up organic participation in the same subreddits.