ContactsThe engagement score
A number you can argue with. Your rules, your decay rates, and a decayed sum of things that actually happened, fed by every other product you run.
You write the rules. Every one of them.
An activity, the points it is worth, and how fast those points fade. Points can be negative, decay can be off, and a rule can be switched off without being lost.
- One rule per kind of activity you care about
- The score it is worth, and it can be negative
- A half-life in days, or no decay at all
- Enabled or not, without deleting the rule
- Scored onto a track, so giving and overall can differ
Then it just adds up. With the old stuff weighing less.
Each scored activity is kept as its own fact, carrying the points and half-life it had at the time. The score is those facts, decayed and summed. Nothing else.
- Every scored activity is kept as its own fact
- Carrying the score and half-life it had at the time
- So editing a rule changes what happens next, not what happened
- The score is a decayed sum of those facts, nothing more
- Held to a whole number between 0 and 1000, never outside it
Four words over the number. So a list reads at a glance.
Cold, warm, hot, champion: a plain ladder over the score, with the figure still there when you want to look at how it got there.
- Cold below 10, warm below 50, hot below 150, champion above
- Four words, so a list is readable at a glance
- And the number underneath, when you want to argue with it
Three things worth being clear about. Especially the first one.
It is arithmetic, not AI
A decayed sum of numbers you chose. There is no model, no prediction and no propensity score, which is exactly why you can defend the output to a board.
It cools on its own
A half-life means last year’s enthusiasm stops counting like this year’s, without anybody running a clean-up. Set it to zero where an act should never fade: a gift, an attendance.
Rules aren’t retroactive
Change a rule and future activity scores the new way. Every fact already recorded keeps the weighting it was given, so a tweak in June cannot silently restate your January numbers.
The rules, and what they produce
Eight rows of settings on one side, and a number on the other that anybody in the room can trace back to them.
The score is the signal. Acting on it is a workflow.
You do. Every rule is a row you write: the activity, the points, the decay. There is no vendor’s idea of engagement underneath it, which is the point: a score you cannot explain is a score you cannot act on.
No. It is a decayed sum of the scores you assigned to things that actually happened. No model, no propensity, no machine learning. If somebody asks how a contact got to 186, the answer is a list of facts and arithmetic.
Yes. A score delta is a signed integer, so an unsubscribe or a bounced address can cost points rather than merely failing to add any.
No, deliberately. Each scored fact keeps the weighting it had when it was recorded. Retroactive rescoring would mean your history changed every time somebody adjusted a setting, so it does not happen.
From whichever Mercleo products you run. Pageviews, email opens and clicks arrive from Pulse, attendance from Engage, applications from Cadre, and gifts and deals from Reach itself. Nobody types any of it in, and with Reach alone the score reads the gifts, deals and stage changes it can see.
Yes: crossing an engagement threshold is one of the workflow triggers. The scoring engine produces the number; what to do about it is a workflow you build.