Flagship engagement
Retention and cohort reporting engagement
A time-bounded engagement that reconstructs named customer groups, builds period-by-period return tables, and leaves a written briefing your operators can reuse.
Who it is for
Product, membership, and customer-experience leads who already keep order or visit histories and need those histories read as cohorts rather than as a single blended average.
What you leave with
A dated briefing pack: group definitions, a retention table for the agreed window, annotations for known events (price changes, closures, campaigns), and a live working session to walk the numbers with the people who will use them.
Scope
One business line or program, one geography if records are split that way, and a reporting window of 8 to 52 weeks depending on how often customers typically return.
Included
- A written definition of each cohort (join week, first purchase, first visit, or another event you name)
- Cleaning notes for missing dates, duplicate records, and groups too small to interpret
- Period-by-period return tables and a short narrative of what the table does and does not show
- One 90-minute interpretation session by phone or in Duluth by arrangement
- A revision pass if source files change before the agreed freeze date
Outside this engagement
- Live connections to payment processors, wallets, or exchanges
- Advice on buying, selling, or holding any asset
- Staffing a standing analytics desk inside your organization
- Software installation on your machines as part of the fee
Who does the work
Bundlelogicbase reporting staff in Duluth, working from files you provide. The optional desktop utility is separate and remains a local record-keeping aid.
How the work proceeds
- Intake call to name the customer event that starts a cohort and the return event that counts as retained
- File freeze: you send exports; we log what arrived and what was unusable
- Draft tables with event annotations
- Working session to correct misread operational context
- Final briefing pack in PDF and a plain-text table
Duration
Three to five weeks from file freeze, shorter if the window is under 12 weeks and the extract is already dated.
Where the work happens
Remote by default. In-person sessions can be arranged around Duluth when both sides prefer a printed table on the table.
What to prepare
A dated extract with a stable customer identifier, the starting event, and later activity timestamps. We do not need payment credentials, private keys, or live system access.
Constraints
Groups smaller than an agreed threshold are reported as too thin to interpret rather than dressed up with a percentage. Seasonal businesses must include at least one comparable prior season or the briefing will say so.
How the fee is set
Quoted from the number of cohorts, the length of the window, and whether source files need reconstruction. See the rates page for starting figures.
Next step
Call or email with the program name and the date range you want read. There is no intake form on this site. Contact details · Rates