“Is sales growth coming from more customers or larger baskets?”
Fundle Brain
01
Set the comparison
Choose the question, dates and business scope. Agree whether the comparison is calendar-based, like-for-like or cohort-based.
02
Check the data
Review freshness, missing stores and return treatment. A change in reporting coverage should not be mistaken for a change in trading.
03
Break down the movement
Separate customer count, purchase frequency, basket value and category or store mix. Follow the largest relevant driver into its supporting records.
04
Assign the next investigation
Write a concise finding with evidence, remaining uncertainty and a useful action. Route it to the team that can change the outcome.
BUILT AROUND THE WORK
A number is useful when you can explain its movement.
Fundle’s retail command centre brings city, store and date filters together with sales, basket, repeat behaviour and loyalty measures. Fundle Brain adds a business-question interface so the analysis can begin with what the team needs to understand.
01
Store and customer context
Read sales, customers, transactions, average order value and units per transaction within the selected city, store and period.
02
Retention and loyalty context
Review repeat behaviour, signup cohorts and points liability together. A sales increase driven by new acquisition tells a different story from stronger existing-customer demand.
03
Trends and data health
The command centre includes sales trends and API-health visibility. Check coverage and freshness before interpreting a sudden change.
04
Questions that lead somewhere
Use the diagnosis to frame a campaign, store investigation or merchandising discussion. Carry the period, filters and metric definition into the resulting brief.
THE OPERATING FLOW
From the first question to the next useful action.
Why can sales rise while the repeat-purchase rate falls?
Sales can increase because the business attracted more new customers, sold larger baskets or changed its store and product mix, even while a smaller share of customers returned. Compare consistent customer cohorts and store sets, then separate customer count, purchase frequency and basket value. Confirm the repeat-rate denominator, observation window and return treatment before deciding that retention has deteriorated.
01
Set the comparison
Choose the question, dates and business scope. Agree whether the comparison is calendar-based, like-for-like or cohort-based.
02
Check the data
Review freshness, missing stores and return treatment. A change in reporting coverage should not be mistaken for a change in trading.
03
Break down the movement
Separate customer count, purchase frequency, basket value and category or store mix. Follow the largest relevant driver into its supporting records.
04
Assign the next investigation
Write a concise finding with evidence, remaining uncertainty and a useful action. Route it to the team that can change the outcome.
BRING YOUR BUSINESS QUESTION
Give Fundle Brain something useful to work on.
Start with the question. Define the data and workflow needed to answer it.
↗Is sales growth coming from more customers or larger baskets?
↗Which stores are losing repeat customers despite higher overall revenue?
↗Does the apparent decline remain after matching the store set?
CONNECT
What data is needed for reliable AI retail analytics?
Reliable retail analysis starts with the decision and a consistent dataset, not with the AI model. For sales and store comparisons, use transaction dates, amounts, returns, product definitions and a stable store hierarchy. For repeat-customer analysis, add a consistent customer or member key where appropriate.
Before comparing results, check which stores and periods are covered, how late records and refunds are handled, and when each source was last updated. Agree the metric, denominator and comparison window. A missing store feed or a changed customer definition can look like a trading change. Use only the data needed for the question, and check material findings against the source records.
PUT IT TO WORK
Leave with a useful output.
Conversational analysis, supporting comparisons and prioritised action briefs.
MEASURE WHAT MATTERS
Repeat rate / like-for-like sales / basket value
Use stable metric definitions and comparable periods. A useful answer exposes the scope and supporting comparison; faster reporting is valuable only when the answer remains reproducible.
GOOD QUESTIONS
Before we get started.
Can I trust every AI answer without checking?
Use the underlying period, filters and metric definitions to assess the answer. Material decisions should be checked against source data and the agreed business rules.
What can the team explore in the command centre?
Use city, store and date filters to explore sales, average order value, repeat behaviour, customers, transactions, units per transaction and points liability. Sales trends, signup cohorts and API health provide additional operating context.
What does Fundle Brain add to a dashboard?
It provides a business-language starting point for investigation. The interface supports questions across loyalty, sales, customers and campaigns; the underlying records and definitions still determine the quality of the answer.
