APAI Praveen
← All projects
MARKETINGSTARTER~60 minFREE SAMPLE

Keep a diary of your own impulse buys and what set each one off

Most unplanned purchases feel like free choices, yet a banner, a friend or a mood usually nudged them along.

WHAT YOU HAND IN

A one-page diary table of your last ten unplanned purchases, with the trigger, the mood and whether you were glad afterwards, plus a short note on your top two triggers.

YOUR AI BUILD GUIDE

This is a field study project. Collect your own evidence (a survey, interviews, a diary or a visit) and find the pattern in it.

The approach: Decide what you need to learn before writing a single question → Collect from real people or real places → Code the answers into themes → Report the pattern and how sure you can be.

Think first

This is a study of one buyer: you. The real question is what tips a purchase from 'no plan' to 'paid', so you are looking for the moment just before each buy, not judging whether the buy was sensible.

ANSWER THESE BEFORE YOU OPEN ANY AI TOOL
  • What counts as unplanned for you: not on a list, or not intended that morning?
  • Over what period will you look back, so the ten purchases are not all from one festival week?
  • Will you include small spends like snacks and recharges, or only things you would notice on a statement?
  • How will you tell a trigger (what happened) from a reason (what you tell yourself afterwards)?
WHAT GOOD LOOKS LIKE

A tidy table of ten real purchases, each with a trigger you can name in a few words, and a short note on two triggers that show up again and again, with an honest line on whether you were glad you bought.

THE COMMON WRONG TURN

Writing the trigger you think a marketer would expect ('it was on discount') instead of what you actually saw or felt. Your memory will tidy the story; write it down before you explain it.

Gather

COLLECT
  • Your UPI history, card statement or wallet app for the period you chose
  • Order histories in the shopping and food apps you use
  • Any screenshots, notifications or messages from just before a purchase
WHAT AI MUST NOT MAKE UP

Do not let AI suggest purchases or triggers for you. Every row must be a purchase you made and a moment you remember; if you cannot remember the trigger, write 'not sure' rather than a guess.

Build with AI

  1. 3.1 Set up the diary tableCLAUDE CHAT

    Why this tool: A quick conversation gets you a sensible set of columns before you start filling anything in.

    Ask Claude for column headings only, then copy them into a sheet or a table in a document. Keep the columns few so you actually fill them.

    Prompt to copy · fill in the [brackets]
    I am keeping a diary of my last ten unplanned purchases to find what triggers them. Suggest five or six column headings for a simple table, including one for the trigger, one for my mood and one for whether I was glad afterwards. Explain each heading in one line. My purchases are mostly [the kinds of things you buy, e.g. food delivery and clothes].
  2. 3.2 Fill in ten rows from memory and recordsGOOGLE SHEETS

    Why this tool: A sheet lets you sort and count rows later without retyping anything.

    Go through your records and add ten unplanned purchases. Write the trigger in your own words first, in the moment you remember it, before asking AI anything.

  3. 3.3 Group the triggers into typesCLAUDE CHAT

    Why this tool: Claude is good at proposing a grouping; you are the only one who can say whether it is true.

    Paste your trigger column only (no amounts are needed) and ask for a grouping. Then go row by row and move anything that was grouped wrongly.

    Prompt to copy · fill in the [brackets]
    Here are the triggers I wrote for ten unplanned purchases, one per line:
    [paste your trigger column]
    Group them into three to five types, such as discount, social pressure or boredom. Show which line goes in which type. Where a line could fit two types, say so instead of choosing for me.
  4. 3.4 Draft the note on your top two triggersCLAUDE CHAT

    Why this tool: Asking for questions instead of a finished note keeps the conclusions yours.

    Count which types appear most, then ask Claude to interview you about the top two before you write the note yourself.

    Prompt to copy · fill in the [brackets]
    My two most common purchase triggers were [type one] and [type two]. Ask me five short questions, one at a time, that would help me write an honest paragraph about why these two work on me and what I could do differently. Do not write the paragraph for me.

Check the AI's work

  • Every row is a real purchase you can find in your records
  • Each trigger describes what happened before the purchase, not a reason invented afterwards
  • The grouping still matches what you remember for every row, after you moved the wrong ones
  • Your note says honestly how many of the ten you were glad about
  • Read a sample of raw answers yourself before trusting any grouping
  • Say how many people or observations the finding rests on
WHAT AI USUALLY GETS WRONG HERE

AI tends to file almost everything under 'discount' or 'convenience' because those are the textbook answers. If your groups look like a textbook list, go back to your own words in the trigger column.

Finish and show

DONE WHEN
  • Ten rows, all real, with trigger, mood and glad-or-not filled in
  • Three to five trigger types, each with its rows
  • A short note on your top two triggers in your own words
  • Amounts removed or rounded if you will share it
HOW TO PRESENT IT

One page: the table on top, the note underneath. A reader should see your two main triggers within ten seconds.

FOR YOUR PORTFOLIO

In your write-up, explain one trigger you did not expect and what it taught you about how marketing reaches people like you.

Which tool fits

  • Claude chat: A quick conversation gets you a sensible set of columns before you start filling anything in.
    Free plan works; Pro gives longer conversations.
  • Google Sheets (free alternative): A sheet lets you sort and count rows later without retyping anything.
    Free with a Google account.
  • Google Forms (free alternative): a free survey that fills a sheet for you
    Free with a Google account.
LEVEL UP

Ask three friends to keep the same diary for a week, then compare whose triggers differ and why. That turns a personal log into a small field study.

AI helps you think, check and polish. The work, and the judgement, stay yours.

Like this one? There are many more.

This is a free sample. Full Access opens every project in full and lets you submit your finished work.

See Full Access →
More starter projects in Marketing
MARKETINGSTARTER
Interview a family member about how they chose their last phone

Brands spend heavily on ads, but real buyers often decide because of a cousin's advice or a shopkeeper's push.

~75 min🔒 Details
MARKETINGSTARTER
Compare the voice of three snack brands in their social posts

Two brands can sell similar biscuits yet sound completely different, and that voice quietly shapes who feels welcome.

~75 min🔒 Details
MARKETINGSTARTER
Audit the social media profile of a small local business

Many small shops have a profile that is unclear, outdated or hard to contact, and they lose customers without knowing it.

~75 min🔒 Details