Intent First: Designing SEO Machines: Lessons from Real Projects

Discover why intent – not keywords – is the real engine behind scalable SEO systems, with actionable frameworks drawn from real client and startup projects that actually work.

Transcript

1. Intent First: Designing SEO Machines* *not actually machines ;) Yagmur Simsek Founder, Search 'n Stuff Athens SEO 2026
2. About Me Athens SEO 2026
3. Athens SEO 2026
4. Let's talk about the laundry. Two days. A lot of machines. But now;
5. We throw everything in one basket: keywords, every signal, every rule.
6. Intent is how you sort it. Keywords are the laundry, not the machine.
7. SORT By intent and persona. Never wash silk with denim ;) WASH The right cycle for this client - a custom framework. FOLD Clean, intent-matched pages. Then measure. The intent wash cycle
8. AI integrations. Automations. Vendors, SaaS, agent stacks. Brilliant experiments - mostly on giant sites.
9. But some of us have a few big projects, and a lot of mid-size and small clients.
10. Same craft. A different machine for every client.
11. • You want SEO actually shipped • So you can test, learn, grow • That needs the client's buy-in • Buy-in follows their priorities • Not just your audit Why bend the strategy?
12. Find the sweet spot - between your SEO plan and businesses’ priorities for Q1-Q4, H1, H2.
13. How I actually start with a client.
14. Before the audit UNDERSTAND A few meetings on their business - not their keywords. EMBED Join team days & quarterly updates; take notes. ALIGN Tie every SEO play to their H1 goals.
15. • Sales & revenue targets • Product & collection pages • Distributors & offline events • Activation campaigns • Legal, finance, CX - all talking Learn the new launch at kickoff - not in month 3
16. Same audit, same plan for everyone. You arrive late and do the work twice. GENERIC ONLY SEO-LED Built around their priorities. It ships, gets tested, grows the business. SEO + CLIENT TEAM-LED
17. Then it becomes a machine FRAME Build a format that fits this client. REPEAT Run it again next quarter. AUTOMATE Now pick the tool to do it for you.
18. Automate the repeat - not the thinking.
19. • Privacy blocks you? Have their team runs it • Beyond your skill? Bring the right person • Your 10 hours vs their 3 hours • Collaborate - don't hero it Know your limits. Protect your time.
20. The real engine isn't keywords. It's buyer states.
21. Who is the buyer? 'Marketing Mary, 35'. A body, not a buying moment. WRONG QUESTION What are they in the middle of? A persona is a moment. One person is three personas a day. RIGHT QUESTION
22. • Demographics, not moments • Built from imagination • Too few, too broad • Filed away, never re-read Why most personas fail
23. A persona you can write from • The buyer state, and which offering it maps to. WHO • The core mindset and the real pains. WHY • Product angle, content themes, coverage score. WHAT
24. Real personas come from data • GSC queries • Support tickets • Live-chat / chatbot logs • Lost-deal & objection notes • Survey answers • Cart-abandonment & return/refund reasons (ecommerce) • Form fields like "how did you hear about us? OWNED • Reviews • Sales-call transcripts • Social media comments • Customer & win/loss interviews • Your own community (Slack / Discord / WhatsApp groups) • Testimonials & case-study call recordings • App-store reviews • Webinar / event Q&A VOICE • Reddit. Quora forums • AI-engine prompts • Search autocomplete & "related searches" • People Also Ask (AlsoAsked) • Competitor reviews: mine their 1–3 reviews; every complaint is a gap you can own THE WILD
25. 14 / 78 Why most personas fail? If you can't write a content title from your persona, it's not a persona. FAIL 1 "Marketing Mary, 35" Demographic descriptors describe a body, not a buying moment. Useless for a content strategist. FAIL 2 Built from imagination Written in a meeting, never cross-checked with search data, reviews, or sales transcripts. FAIL 3 Too few, too broad One persona per segment is a starting point, not a strategy. FAIL 4 Never re-read Built once, filed away, never reopened when briefs are written. Lives in the deck, not the work.
26. 15 / 78 A persona is a buyer state, not a person. One human can be three different personas in one day. WRONG QUESTION Who is the buyer? Leads to demographic cardboard cutouts that don't drive content decisions. RIGHT QUESTION What are they in the middle of? Leads to specific moments, problems, and language you can actually write to.
27. 15 / 78 A persona is a buyer state, not a person. WRONG QUESTION Who is the buyer? A 20-year-old student. A demographic you can't write to. RIGHT QUESTION What are they in the middle of? Trying to study at 11pm — tired, running on caffeine, and knowing they’ll crash at 2am. > Coffee Refugee at 7am > Afternoon Slumper at 2pm > Late-Night Studier at 11pm.
28. Prompt > publish. More content, same yield. No point of view. AI-GENERATED AI sorts at scale; humans decide. Checked against GSC and e.g. the top 10. AI-INFORMED
29. 16 / 78 TEMPLATE The seven fields that actually matter. If you can fill all seven, your writers can brief themselves. 1 Description Concrete sketch of the buying moment — who, what, when. 2 Product / Service category fit Which of your SKUs / offerings this state maps to. 3 Core mindset The one-sentence inner monologue in this state. 4 Key pain points The specific frictions that make them search / scroll. 5 Product angle The benefit framing that lands for this state (not the feature list). 6 Content themes 3–5 clusters of titles this persona will click and read to the end. 7 Coverage assessment How well your current site already serves this state — colour-coded.
30. Real projects. Modest results - but frameworks that travel.
31. US client selling caffeine gums, mints, energy gums etc.
32. Score coverage in colour GREEN Strong coverage - minor updates only YELLOW Partial - expand the existing page ORANGE Weak - significant rewrite RED Minimal - new or major rewrite BLUE No page yet - create it PURPLE Exists, but off-plan - reintegrate
33. • Functional product, lifestyle audience, one product line, 25+ buyer states. • We mapped content to the moment - the 2pm slump, the gamer's crash - not to 'caffeine gum'. US client selling caffeine gums, mints, energy gums etc.
34. 17 / 78 25+ buyer states (personas) 596 content titles planned 514 with existing blog URL 82 new pieces still to write US client selling caffeine gums, mints, energy gums etc.
35. US client selling caffeine gums, mints, energy gums etc. Then they decided to change the Shopify template and we had to pause content uploads…
36. US client selling caffeine gums, mints, energy gums etc. While waiting, we used the persona and content sheet for the link equation strategy roadmap
37. US client selling caffeine gums, mints, energy gums etc. The first article is finally published.
38. UK Client in share plan, reward, financial reward communication consultancy
39. UK Client in share plan, reward, financial reward communication consultancy
40. • AI and organic visibility growth started from month 3 • Built stronger trust in long-term SEO value • Expanded client buy-in for editorial growth • In a niche low-volume market, every click and impression mattered for brand’s visibility UK Client in share plan, reward, financial reward communication consultancy
41. UK Client in share plan, reward, financial reward communication consultancy
42. Global Client with cyber security products/solutions
43. • B2B deep tech: low volume, long cycles, technical buyers. • We built content around technical-validation intent, the real questions enterprise buyers ask internally [ Add your screenshot here ] Global Client with cyber security products/solutions
44. • Alongside the content strategy, we also invested in; ◦ niche editorial placements ◦ industry-specific publisher collaborations • These off-site articles linked back to the related campaign / landing pages, helping strengthen the overall ecosystem around the topic. Global Client with cyber security products/solutions
45. Search 'n Stuff, Crowdfunding Campaign: The trend that built itself We launched Antalya as a 'Crowdfunding Edition' that runs only if 50 tickets sell by Jan 1. THE PLAY Built for emotion, belonging, urgency, ownership. We read the audience, not the keyword. THE INTENT Sessions spiked; referring domains went ~2 to 30+, almost all organic. I barely promoted it - the awareness stuck. THE TREND
46. Optimise the machine that turns human intent into architecture. Stop optimising for search engines.
47. If you remember three things INTENT FRAMEW ORK FIT over keywords over budgets client + SEO-first, not one-size Read the human. Build the machine around them.
48. SEO, AI, PR tools I use recently SEO & CRAWL Ahrefs · Semrush · Screaming Frog · Majestic · Advanced Web Ranking AI & GEO Claude Cowork · ChatGPT · Gemini · NotebookLM · GeoGenie · Waikay · Convomind.ai GROUND TRUTH Google Search Console · GA4 · AlsoAsked AUTOMATE & AMPLIFY n8n automations · Whitepress (links) · Finchling The stack and method changes per client.
49. 15 / 78 Remember! “ The best time to plant a tree was twenty years ago. The second best time is now. Chinese Proverb
50. Thank You! Yagmur Simsek Founder, Search 'n Stuff yagmur@searchnstuff.co.uk searchnstuff.co.uk Let's connect > As a small thank you for your time 😊 You can use the promo code below to get 40% off your Search ‘n Stuff London Conference ticket. ATHENS40