Predicted customer value
Early in the customer's life.
SOLUTION
Know what each customer is worth, and tell every platform.
Your order and CRM history already separate your best customers from one-time buyers. Datahash turns that history into predicted value, value bands, and lifecycle stages, then feeds them to bidding and audiences on Meta, Google, and the other platforms you buy on.
IN ONE SENTENCE
Customer value and segmentation is a solution that scores each customer from first-party order and CRM history, predicted value, value band, lifecycle stage, and category affinity, and sends those scores to ad platforms as conversion values and audiences.
WHEN THIS IS YOU
The signs teams bring to us.
- Your top customers are worth many times your average, and your campaigns treat everyone the same.
- Acquisition ads keep reaching people who already buy from you.
- Lookalikes are seeded from every buyer instead of your best ones.
- Your data team has customer scores that never leave the warehouse.
What we compute.
Value bands
From recency, frequency, and spend.
Lifecycle stage
Prospect, first buyer, repeat, lapsing, lapsed.
Existing-customer status
So acquisition spend can exclude or down-weight them.
Category affinity
So audiences match what each customer is likely to buy next.
Where it goes.
Predicted lifetime value on Meta
Learn morePredicted value as conversion values on Google
Learn moreNew-customer goals on Meta
Custom events for first-time purchasers.
Learn moreNew-customer goals on Google
Google's new customer acquisition goal.
Learn moreAudiences on every platform
High-value seeds for lookalikes, existing-customer exclusions, and stage-based segments, refreshed from your source of truth.
Value bands and lifecycle stages, to bids and audiences.
Illustrative. No client numbers.
How an engagement runs.
-
Discover
Order and CRM history, identifiers, and consent checked; the business question agreed.
-
Scope
Which scores and segments, which platforms, and the test agreed, with a go or no-go on fit.
-
Execute
Scores computed and delivered as values and audiences, verified in each platform.
-
Measure
Results read against your current campaigns with a holdout, then refreshed on a schedule.
Questions your team will ask.
How much history do we need?
It depends on the score. Value bands and lifecycle stages work from a year or so of orders. Predicted value and category affinity need more customers and more history before they beat simpler rules, so we start with what your data supports today and move to predictive scores as the history grows.
Do we need a data science team?
No. Our consultants work out with you which scores the goal needs and what it takes to produce them from the data you hold. If your team already has customer scores, we can use them. Bring whoever owns your customer data to the first working session.
Does customer-level data leave our control?
Scores reach the platforms the same way conversions do: hashed identifiers on consented records, sent server-side. The platforms receive a value or an audience membership, not your customer database or the logic behind the score. Consent is checked before any record is sent.
Can we stop paying to reach existing customers?
Yes. Existing-customer and recent-buyer exclusions are among the first things we set up, because they need no prediction at all, just an accurate, refreshed list of who already buys. Acquisition budgets then go to people who are actually new, and the list keeps refreshing as customers buy.
Adjacent moves on the same stack.
NEXT STEP