INSIGHT CONSOLE
LOCAL --:--:--
3 min read

Execution Isn't Strategy: Upgrading Two Small Notion Products

What the same lesson from my off-page SEO system looked like applied to a grocery price tracker and a fuel mileage log.

I built a strategic layer into an off-page SEO system after a review pointed out it tracked activity well but couldn't say whether any of that activity was worth doing. The same week, I looked back at two much smaller personal tools I'd built — PantryBuddy and Pumpdude — and noticed they had exactly the same gap, just in miniature.

Two tools, one shape of problem

PantryBuddy compares grocery prices across four delivery platforms and tells you which one is cheapest for each item right now. Pumpdude logs every fuel fill-up and calculates what that tank actually delivered in kilometers per liter, after one car getting 15 km/l from its usual pump and 9 km/l from a different one — same car, same week — made the gap too obvious to ignore.

Both worked. Both also had the same blind spot: they told you what was true right now and stopped there. A single price check can't tell you if "cheapest platform" is a real pattern or a one-day fluke. A chart you have to eyeball for outliers isn't actually catching anything — it's asking you to do the noticing yourself, which defeats the point of building the tool.

What changed in PantryBuddy

  • Price history instead of a snapshot. The original version overwrote one price field per platform every time you checked. Now every check gets logged, so "cheapest platform" is based on a pattern, not a single lucky day.
  • Free-delivery thresholds, tracked explicitly. The real savings leak in grocery delivery usually isn't the per-item price gap — it's paying four separate delivery fees because nobody's tracking which platform is close to its free-delivery minimum. That's now a real number you can check your cart against before you check out.
  • A savings log. "Potential savings" on any one item is a theoretical number. Logging actual monthly spend against it is what tells you whether the system is really saving money or just producing a plausible-looking column.

What changed in Pumpdude

  • Stations became their own record, not a label. Every fill-up used to carry a station name as plain text. Now each station is a tracked entity with its own rolled-up visit count, average mileage, and average price — a scorecard instead of a chart you have to squint at.
  • A fixed baseline instead of a moving average. Catching a bad tank by eye, from a chart, is exactly the kind of thing that should be automatic. Every fill-up now gets flagged if it comes in more than 15% below that vehicle's baseline mileage — a number you set once, not one that drifts with every new fill-up (including the bad one you're trying to catch).
  • A monthly budget. One bad tank is noise. A cost-per-km that creeps upward for three months straight is a pattern — bad pump, overdue maintenance, a driving habit — and a budget line is what surfaces that before it's three months of wasted money.

The part that generalizes

Neither of these tools got harder to use. Logging a fill-up still takes the same five fields it always did. What changed is what the system does with that log afterward — whether it just stores what happened, or tells you something you'd have missed otherwise.

That's the same question worth asking about any tracker before calling it finished: does it just record activity, or does it eventually tell you if the activity was worth it?

Join the dispatch

New GA4 breakdowns, career notes, and healthcare-analytics posts — straight to your inbox, no spam.