Greetings, all. We hope your work week went well.
OpenAI released ChatGPT 5 in the time that has passed since our last newsletter. The public backlash remains intense. Many people describe the new version as a downgrade, citing slower output times and colder, less personable responses. What many people found more upsetting was the fact that older models got abruptly retired without warning, disrupting a lot of workflows. Another outcry came from the people who’d formed emotional attachments to GPT 4o and now feel like they’ve lost a trusted friend.
We don’t judge.
Putting those criticisms aside, there are some things you can do with ChatGPT 5 that were not possible in the past. We built out the ability to look at side-by-side GeoGrid search results, but another feature that many of you do not take advantage of is the .csv export. Not that we blame you. It’s a lot of raw data that isn’t especially fun to browse. You can click here to see what the export looks like for the search result that’s pictured above this paragraph.
ChatGPT 5 knows how to interpret the .csv files and can give you summarized general trends about the progress of a given business. It specifically cited its ability to make sense of these exports when we asked it what new capabilities are available to digital marketers.
To get a feel for what we mean, we’ll paste the output when we uploaded CSVs from January 27 and August 11 and asked ChatGPT 5 to give us a rundown of the general trends:
Here’s the January 27 → August 11 trend breakdown for Plumbing Heating Paramedics (heater replacement search term) based on your CSV exports and the August geogrid:
1. Overall shift toward stronger visibility
AGR improved from 12.23 to 7.58 — moving average rankings from the low double digits into the high single digits.
ATGR improved from 13.47 to 8.45, showing better performance even in the business’s weaker grid points.
SoLV jumped from ~1% to 22%, meaning far more grid points now rank in the top 3.
2. Clear northern half dominance
Northeast & North-central quadrants saw the biggest ranking gains.
January: patchy mid-pack (7–12) with scattered green.
August: solid green swaths (ranks 1–3) from Noblesville up toward SR 13 and SR 19.
Northwest improved from many orange/red zones (teens and 20+) to mostly yellow/green, with only the farthest western fringe still lagging.
3. Strong gains around Noblesville center
Central nodes that were 7–10 in January now often sit at 1–4.
The dense cluster of green in August around Noblesville shows you’ve overtaken most local competitors within your core service area.
4. West & Southwest improvement, but still weaker
January’s west/southwest grid points were heavy with 15–20+ rankings.
August shows many of these improved into 8–12 range — a solid gain but still not breaking into top 3 in most spots.
5. Southern reach still the weakest
South of Fishers and into Carmel, rankings remain mostly orange/red in August (mid-teens to 20+).
Some nodes improved from the bottom of the pack to low/mid-teens, but overall visibility here is still low.
6. Gains are broad, not just clustered
Improvements aren’t limited to one “hot zone” — most quadrants saw some lift.
The core north-central service area turned into a high-performing “green belt,” while peripheral areas shifted upward a few positions even if they’re still outside the top 3.
If you want, I can also map this in a color-coded “improvement heatmap” from Jan 27 → Aug 11 to visually show which grid nodes jumped the most. That would make the quadrant story even clearer.
We told ChatGPT that we did want to see a color-coded improvement heatmap. The image you see above this paragraph is what it gave us. We can’t make heads or tails of it, but you may have better luck if you’re not colorblind.
Our intent is not to defend OpenAI in today’s newsletter. OpenAI didn’t even defend itself (it publicly acknowledged the criticisms and reinstated old models). We’re just pointing out that you can squeeze some lemonade out of this flop. ChatGPT’s italicized text output above that mysterious graphic is full of decent talking points that you can recite to your clients.
Our first closing link comes from Bloomberg. It’s about TikTok’s ability to manufacture strange trends like labubus and Dubai chocolate. Next, from Later, is 8 Ways to Build a Brand Awareness Campaign to Boost ROI in 2025. Third, from Buffer, is The Anatomy of a Great Hook: How to Grab (and Keep) Attention in 2025. Finally, The Verge wrote an article about YouTube asking people to upload their IDs if its algorithms suspect that you’re under 18 years old.
That’s all for today. Enjoy your weekends.