My first post about the Monthly Budget Review from Raw Transaction Data prompt was basically a love letter to catching a forgotten subscription. This one is the advanced move: what happens when you stop treating each month as a standalone snapshot and start treating twelve of them as a single, slightly embarrassing story about yourself. The prompt is deliberately strict about single-month trends — if you only feed it one month, it tells you flat out that multi-month trends cannot be assessed. That honesty is exactly the constraint that finally pushed me to do it properly.
Why a single month lies to you
A single month of budget data is a photograph. It is true, but it has no motion. You cannot tell from one frame whether your marketing overspend was a one-time annual invoice or a creeping habit, and you cannot tell whether your 'great' low-spend month was discipline or just a quiet billing cycle that is about to catch up with you. I kept making decisions off single frames and then being confused when the movie did not match.
The setup: same prompt, twelve runs, one summary tab
Each month I run the prompt on GPT 5.2 with that month's transactions and that month's budget targets. Nothing fancy. But instead of reading the output and closing the tab, I now copy two fields out of every single run into a plain spreadsheet. Twelve rows later, I am no longer looking at a month — I am looking at a trend the single-month run is explicitly forbidden from claiming on its own.
- Column A: the month
- Column B: net over/under percentage, pulled from the Summary block
- Column C: the number-one biggest-swing category that month, from Biggest Swings
- Column D: one sentence pasted straight from the prompt's likely-driver note
- Column E: did I act on the recommended action that month — a simple yes or no
What twelve months revealed that one month hid
The first pattern that jumped out was seasonal and genuinely embarrassing. My 'software' category appeared in the Biggest Swings list in four separate months — and every single time it was a different annual tool auto-renewing on a date I had forgotten. I do not have a spending problem. I have a renewal-date blindness problem, which is a much more fixable diagnosis. One month's review made each renewal look like an innocent one-off accident. Twelve months stacked together made it a clear system failure, which I fixed in an afternoon by dumping every renewal date into a single shared calendar.
The second pattern was kinder to me, but still revealing. My net over/under percentage was reliably worst in the two months right after I closed a big client, because I would 'celebrate' with new spend before the cash had actually landed in the account. Seeing that show up in column B four times across the year was the financial equivalent of a fitness tracker quietly logging your 11pm snacks. You cannot argue with your own data when it is sitting in a row with a date next to it.
There is a third, subtler payoff I did not expect: the twelve-month view changed how I read each individual monthly run. Once I had the trend tab, a single bad month stopped triggering panic, because I could see instantly whether it was an outlier or the start of a slope. Context turns alarming numbers into informative ones. A 30% marketing overspend reads very differently when column D already tells you it is the annual conference invoice that lands every June, versus when it is a brand-new line you have never seen before. The prompt gives you the month; the tab gives you the meaning.
The honest caveat
I will not oversell this. The prompt does not do the stitching for you — it is a per-month analyst, and it correctly refuses to fabricate a yearlong trend from a single input. The twelve-month view is manual, maybe fifteen minutes a month of copy-paste into the summary tab. But the discipline is the entire point. The strictness of the prompt is what forced me into a real, repeating review habit rather than a once-a-year panic when taxes were due.
If you already run the Monthly Budget Review from Raw Transaction Data prompt on Prompt Dock, try keeping the two-field summary tab for a single quarter and see what surfaces. Pair it with the Simple Runway and Burn Rate Explainer prompt at the same time, and you will have both 'where did it go' and 'how long does it last' answered in one Monday-morning sitting — which is roughly the entire job of a part-time CFO, done with a coffee and a spreadsheet.