Segmenting a customer database

How can you segment your customer database in a useful and verifiable way with AI?

1960 € per session, excl. VAT

In short

“Segmenting a customer database” is a private in-person session of a full day for marketing, CRM and digital marketing managers or marketing analysts and product managers who use a customer database. It helps you prepare and anonymise a customer export before any assisted analysis. Groups of up to 12 participants, in French, with exercises on your own work cases.

Why this training?

Many customer databases are segmented out of habit (size, sector, region), with no link to purchasing behaviour. Campaigns sent to these segments remain largely irrelevant. Generative AI helps explore an export, suggest groupings and draft segment profiles; the session requires numerical verification and compliance checks to ensure that the segments are real, stable and usable by the team. Running case study: ‘Fleurs & Co’, a fictitious florist with an online shop and two shops, has a fictitious export of 2 000 customers (purchase dates and amounts, channel, categories purchased). The team wants to stop sending the same newsletter to everyone. Transfer to the workplace: the participant applies the approach to a real export prepared with their data department, then tests the planned action for the priority segment and measures its effect.

Objectives of this training

  • Prepare and anonymise a customer export before any assisted analysis
  • Choose behaviour-related segmentation variables (recency, frequency, monetary value, usage)
  • Build four to six segments and verify their size and characteristics through calculation
  • Write an evidence-based profile and priority needs for each segment
  • Define a marketing action and an indicator for each segment
  • Document the allocation rules so that the segmentation is reproducible

Who is it for and what are the prerequisites?

Roles and tasks concerned

  • Marketing, CRM and digital marketing managers
  • Marketing analysts and product managers who use a customer database

Prior skills

  • Regularly use a generative AI assistant and know how to structure a request (AI Prompting level or equivalent skills).
  • Have practical experience of your organisation’s marketing: target audiences, channels, content and key indicators.
  • Know how to use a spreadsheet: filters, sorting and a simple pivot table.

What you will work on

  1. Why and how to segment

    Objective: Know which decisions the segmentation should support. Descriptive or behavioural segmentation; recency-frequency-monetary value approach; intended marketing decisions; data protection rules.

  2. Prepare and anonymise the export

    Objective: Have clean, compliant data. Removing identifiers, grouping rare values, checking duplicates and missing values; what may be sent to the authorised AI tool.

  3. Explore the data with AI

    Objective: Identify distinct behaviours. Exploratory requests: distributions, cross-tabulations, outliers; systematic recalculation in the spreadsheet of the figures reported by AI.

  4. Build and test the segments

    Objective: Obtain distinct, sufficiently large and stable segments. Explicit allocation rules; minimum sizes; testing across two periods; merging or removing segments that are too similar.

  5. Segment profiles

    Objective: Make the segments understandable to the team. Evidence-based profile: quantified behaviour, likely needs, channels; distinction between figures and hypotheses; neutral naming without stereotypes.

  6. Action plan by segment

    Objective: Move from analysis to campaign. One action, one message, one channel and one indicator per segment; prioritisation based on value and effort; update frequency.

What you take away

  • Export preparation and anonymisation log
  • Segment table with allocation rules, sizes and stability
  • Evidence-based segment profiles
  • Marketing action plan by segment with indicators

Method and assessment

  • Short teaching inputs followed by immediate practice on the running case study.
  • Individual work followed by work in pairs, with peer review of the outputs.
  • Guided group review: what works, what needs correcting and what remains a human decision.
  • Preparation: Identify a possible customer export with your data department, without names, email addresses or direct identifiers. List last year’s campaigns and their target audiences.

Observable criteria: No direct identifiers appear in the data sent to the tool; Every key figure reported by AI is recalculated in the spreadsheet; Each segment has a reproducible allocation rule and a documented size; The profiles distinguish quantified facts from hypotheses; Each segment is linked to an action and an indicator.

Frequently asked questions

Can we upload our customer database to the AI tool?

Only an anonymised export, and only in a tool authorised by your organisation; the session shows how to prepare this file.

Are statistical skills required?

No. The method relies on simple rules and checks in a spreadsheet, without an advanced statistical model.

Does AI do the segmentation for us?

It suggests groupings and speeds up the analysis; the final rules are decided and verified by the team.

How many segments should we aim for?

Generally four to six: enough to tailor messages, but few enough to be used effectively.

Last updated: