Jess Wright

@jess.writes

Spent 6 years doing content marketing for other people's SaaS, finally building my own. Case studies are my love language.

Last 30 days12 events

Posts (3)

📈 Wins & losses (case studies)

AI Overviews ate our long-tail content. Out of 4,000 posts, 12 got cited. Here's what they had in common.

Did the post mortem nobody wants to do. We ran an audit of 4 years of content — roughly 4,000 published posts on the blog. Pulled every AI…

Jess Wrighttoday
320
📈 Wins & losses (case studies)

One original survey = Perplexity citations in 3 weeks. Here's the exact format that worked.

Background: i've done content marketing for other people's SaaS for years, finally building my own tool, so i know the playbook. Decided to…

Jess Wright2 weeks ago
3441
📈 Wins & losses (case studies)

Paid $7,200 to a GEO agency. Month 6 postmortem.

I want to write this up while the sting is still fresh enough to be useful. $1,200/mo for 6 months. Here's the complete list of what they…

Jess Wright2 weeks ago
1830

Recent replies (9)

  1. Adding rule 4 from our own attempts: never link to anything on the first product mention. A bare brand name reads as recommendation; a brand name plus a link reads as marketing. Perplexity seems to weight bare-name mentions higher than linked ones for whatever reason. Sample size of 2 campaigns but the difference was visible.

  2. ngl this changed how i think about content republish cadence. We've been refreshing posts on a fixed quarterly rhythm, but if Perplexity reads weekly and ChatGPT reads monthly, refreshing the 5 Perplexity-cited posts weekly probably moves more pipeline than blanket-refreshing 50 posts quarterly. Going to test that next month.

  3. The training-data lag point is what separates GEO from every previous SEO wave for me. Doing everything right today and seeing nothing for 6-8 weeks is psychologically brutal in a way that SEO never was. At least with SEO you'd see crawl confirmation in hours. The delayed feedback loop makes it really hard to know if you're on the right track.

  4. The defensive brand probe point is something I wish I'd added from day 1. Caught a bad Perplexity answer on 'is [brand] still maintained' 2 months in. Would have been embarrassing if a prospect saw it first.

  5. The ChatGPT 'what category is [brand] in?' probe is a useful quick test. Tried it on my own brand and got back a category description that was adjacent but slightly off. Explains why I'm showing up for some prompts but not others.

  6. The month 7 despair is so real. I'm at month 5 right now with modest movement and this is weirdly reassuring. Not that I should expect a miracle, but that the curve can be non-linear and patience has historical precedent.

  7. This is actually useful for a different problem I have. We changed our product name 14 months ago and ChatGPT still uses the old name. Going to try adding a 'formerly known as X' line directly in the hero body text. Did you place yours above the fold, or is it okay buried lower on the page?

  8. The 9-day Perplexity update after a simple blog post is fast. Did you do anything to accelerate the recrawl (like submitting to IndexNow or pinging Perplexity's crawl mechanism) or did it just happen organically?

  9. What's your cadence on the 2 brand-name probes? I added those as weekly but I'm wondering if monthly is enough since they change slowly.