The data said a bigger first order makes a loyal customer. It was mango season.

It started with a question.

Sneha Kulkarniruns growth at Simcorp Freshin Veloccy

Our theory is that households whose first order is bigger stick around more, so we would show higher-value items like sweets and gourmet foods to first-time users.

AI with access

Access

Experience

none

+6.6 pts repeat within 60 days, big first orders against the rest

With access to the data, the theory holds.

Across 18,105 new households, the ones whose first order was above $1,200 came back within 60 days more often. The gap is large and the interval is tight. An analysis of the question as asked says yes.

AI with experience

Access

Experience

Sneha Kulkarni · runs growthMango season · every yearand more

+0.25 pts the gap once month and category are held the same

With experience of the company, there is no gap.

The gap lived in one category and three months of the year. Veloccy could not explain it from anything it knew, so it asked the person who would. Like for like, a bigger first order makes no difference.

How it happened in Veloccy.

A simulated grocery app with quick delivery. We built it and planted the cause, so we know what the right answer is. These are a few moments from the work.

Simcorp Fresh Arcs GRO-1Do bigger first orders come back more?

Running

What it read What it led to Who is on it

Repeat within 60 daysAverage order valueNew householdsMango season · every yearHome page test live · app team

GRO-1 · Do bigger first orders come back more?Arc · running

Bigger first orders do not repeat moreEveryday first orders: a leadEveryday-first test

Sneha KulkarniGrowthVictor MarshStorefront

The arc for this work, as the team sees it: what Veloccy read, what the work led to, and the thread it happened in.

01It asks the person who knows

Veloccy saw a pattern it could not explain, and asked.

No table says mango season. Sneha answered in a line. Veloccy kept it as a season that returns every year, with her as the source, so nobody has to explain it again, to Veloccy or to the next person who joins.

Simcorp FreshArcsGRO-1

Veloccy

The gap sits in fruits and vegetables, from May to July, and nothing on the calendar explains it. Is this a season, a campaign, or a change in how orders were recorded?

Sneha Kulkarni

that’s mango season. people who buy mangoes come back every week or two

Saved to the BrainMango seasonCalendar · 1 May to 31 Jul, every year · source Sneha Kulkarni

02It answers what you meant

The answer was no, and it came with a better bet.

Sweets and gourmet foods have the biggest first baskets and the lowest repeat, so the plan would have put the least loyal products first. Veloccy also told the approver that the app team had a test live on the same screen, before proposing anything of its own.

Simcorp FreshArcsGRO-1

Sneha Kulkarni

Households whose first order is bigger stick around more. Should we show higher-value items to first-time users?

Started from

Mango season · every yearHome page test live · app team

The answer

Don’t build it. Like for like, a bigger first order repeats 0.25 points more. What comes back is an everyday first order.

Everyday first orders repeat 36.5% against 27.0% for occasional ones. That is a lead, not proof, so the next step is a test.

Every number opens the query behind it

03It uses what it learned

The same afternoon, the season changed a different answer.

Basket size was down on the week. Read against the weeks around the season’s end, the move would have looked ordinary. With the season set apart, Veloccy called it real. Nobody pointed it to the note.

Simcorp FreshBrainMetrics

Metric

Average order value

Owned by Sneha Kulkarni

This week: −3.74%
Usual weekly move: −1.60% to +1.53%
Read as: A real move
Left out of the range: The weeks around mango season’s end

Why access to the data was not enough.

01The data said yes.

A careful analysis of the question as asked confirms the theory. Getting to no took a second question nobody had asked.

02The missing fact was in someone’s head.

Veloccy knew what it did not know, and who to ask.

03It was used again within hours.

A note in a document waits to be read. This changed the next answer without anyone pointing to it.

What stayed in the Brain.

Every one of these is experience the next question starts from, in every team. See the Brain

FSimcorp Fresh

CalendarMango season · every year
The season, every year, with who said so.
LearningA bigger first order does not bring people back
The idea that was dropped, and why.
ChangeEveryday-first test
A test, approved by the person who owns the storefront, with no result yet.

Next use case · Found unasked, in Slack Nobody asked. Veloccy found it, tested the fix, and reported the winner, all in Slack.

Which bet is on your desk this month?

Show us, and we will show you what Veloccy does with it.

Talk to us