MVP Thinking in a Changing Search Environment

As search evolves—driven by AI Overviews, shifting user behaviors, and ever-changing SERPs—SEOs must adapt how they build, test, and learn. In this talk, Vanda Pókecz, Head of SEO & Product Lead and co-host of the SEO Product Unveiled podcast, introduces the power of MVP thinking for modern SEO teams.

Instead of waiting for the perfect tool or large-scale rollout, this mindset encourages launching lean, value-driven experiments to test assumptions, gather directional data, and drive smarter decisions. From structured data tests on a single template to a scrappy MVP for monitoring AI Overviews, the session dives into real examples that show how impactful “small bets” can be.

Transcript

1. Small SEO Bets: MVP Thinking in a Changing Search Environment Vanda Pókecz
2. Head of SEO & Product Lead @ AIPMM Certified Digital Product Manager Co-host - SEO Product Unveiled podcast linkedin.com/in/vanda-pokecz | @vanda_lilla Vanda Pókecz
4. Are we the right fit?
5. Agenda Transitioning into an MVP mindset Experimentation & “unusual” data for SEO Example Main learnings
6. Transitioning into an MVP mindset linkedin.com/in/vanda-pokecz | @vanda_lilla
7. The smallest version of a product or feature that delivers value and allows you to learn. MVP
8. Launch to test assumptions & reduce risk through real feedback Data backed argument = resources
9. Examples What’s the smallest possible change I can make that helps me learn whether a bigger idea is worth pursuing?
10. Testing structured data on just one template
11. Testing LLM support with content creation process for a single content module on small number of pages
12. Building a basic way to track something new before investing in tooling or automation
13. MVPs don’t mean low quality… they mean right sized for learning!
14. An MVP helps you answer Is this worth scaling? Answer with data, not speculation
15. Perfect plans → get enough live to learn something useful
16. Experimentation and “unusual” data for SEO linkedin.com/in/vanda-pokecz | @vanda_lilla
17. Once you embrace MVP thinking How do you know what to test? And how do you measure whether it’s working?
18. SEO Crawl Data KW rankings Backlinks Traffic relevant metrics
19. Useful but often reactive What’s happening after launch…
20. The best “experiments” start earlier - during discovery!
21. Collect insights about User behaviour Preferences Search intent Friction points… Before deciding what to build
22. We need to get curious earlier!
23. What’s the user trying to do here?
24. What result would actually help them?
25. Why might they skip over our result and click something else?
26. Examples of SEO discovery & experimentation tools
27. User testing Run a quick test with real people searching Ask: Which result would you click? Why? Or just observe first then ask…
28. Internal site search or support tickets Pull data to see what are people searching for on your site
29. Heatmaps and scroll tracking Are people engaging or scrolling right past? Helpful even for types of components which otherwise don’t have a direct engagement element associated
30. Small scale content tests A/B or observational testing They give you early data - directional signals - that can help you prioritize and pitch better MVPs
31. Before you build, test the problem Before you scale, test the solution
32. Example AI Overviews monitoring MVP linkedin.com/in/vanda-pokecz | @vanda_lilla
33. Sometimes, the most valuable data work isn’t about what we ship, but how we learn
35. May 2023 Jun 2025 SGE available in Search Labs May 2024 AI Overviews launched for all U.S. users Global rollout (100+ countries) Oct 2024 AI Overviews introduced to select EU countries March 2025
36. Our tools were not ready…
37. Source: @lilyraynyc Based in Europe? Checking with VPN? We were flying blind, faced with a new kind of challenge
38. How do you measure something you can barely see?
39. How do you prove it’s affecting your business-critical queries?
40. MVP Thinking Start small, fast, imperfect…focus on value Instead of waiting for a perfect tool, structured rollout or custom dashboard…
41. The MVP monitoring journey…
42. Research, before monitoring Small scale test in SGE Labs days… 50 queries Manual with VPN
43. Were AI answers showing?
44. Were our domains cited?
45. Did the summaries even make sense for commercial & transactional intent?
46. V1: Manual monitoring With AIOs picking up we had to get more systematic
47. V2: Evolving with tools As tools in our SEO stack started releasing AIO detection, we layered them in - this gave us better automation and scale Tip: before you start looking outside, check what you already have available and whether it meets your needs
48. We continued to run manual checks, we had to validate how reliable the tools were
49. We started asking bigger questions
50. What share of our top tracked queries showed AIOs?
51. How often were we cited?
52. What was the search volume share of AIO-triggering keywords vs. total?
53. What was the word count spread?
55. V3: From manual to semi automated First validated that the process worked then started using formulas etc. to semi-automate Less time collecting data, and more time analyzing patterns
56. V4: From MVP to scaling From spreadsheets to “live system” 1. Validate that the process works 2. Integrate APIs 3. Build out dashboards 4. Introduce new domains/markets 5. Start experimenting with other SERP and LLM tracking capabilities/tools
57. The beauty of MVP thinking for SEO
58. Getting your hands dirty is often the only way to get real, relevant data There’s secondary research available, but if you want to understand your business, your domain, your industry…you can’t wait around
59. Main learnings SEO doesn’t always need to be big, complex or perfectly packaged to drive value linkedin.com/in/vanda-pokecz | @vanda_lilla
60. MVPs are not shortcuts They’re learning tools They give you faster feedback, lower the risk, and help you move forward
61. SEO needs more experimentation Whether it’s user testing, small-scale feature pilots, or scrappy tracking setups…
62. You don’t need a perfect tool to start measuring The AI Overview monitoring journey I shared today started in a spreadsheet with a VPN
63. Data doesn’t always mean dashboards It can be manual notes, directional signals, or even screenshots What matters is how you use it to make better decisions, and to tell a story others will care about
64. I challenge you!
65. What’s one thing you’ve been waiting to test?
66. What could you build this week - not to launch, but just to learn?
67. Start there Build your MVP And let your data do the talking!
68. Thank you! Vanda Pókecz Head of SEO & Product Lead linkedin.com/in/vanda-pokecz | @vanda_lilla |