Where not to spend.
A channel score on each person shows who responds to a channel and who leans away from it. The people below the 0.25 line are the ones a plan can stop paying to reach there.
Media plans are usually set from audience averages and reach flags. A flag says a person can be reached on a channel. It says nothing about whether that person responds there. Intelafy scores each person on each channel on the same 0-to-1 scale as every other attribute, with 0.25 as the no-signal line: scores above it lean toward response, and scores below it lean away. Once each person carries a channel score, the waste inside a channel that looks good on average shows up, and so do the responsive people inside a channel that looks poor.
A reach flag answers a different question.
A flag says whether a person can be reached on a channel. A person flagged on five channels can be reached five ways, and the flags give no reason to prefer any one of them. The plan still has to choose, and without a score it chooses from averages.
A channel score answers the planner's question: where is this person worth the spend? It is measured the same way as every other Intelafy score (Paper 01), so it can sit in the same file as the audience, the motivations, and the creative cues (Paper 03).
What 0.25 means on a channel.
Every Intelafy score sits on one 0-to-1 scale (Paper 02). At 0.25 the evidence is neutral. Above it, the evidence leans toward the person responding to advertising on that channel. Below it, the evidence leans toward the person avoiding it.
Each person on the file carries a score for each channel. Some channel data is recorded for the person and some for the household, so where a channel's input is household-level, people who live together carry the same score for that channel.
Because the scale is shared, a 0.72 for connected TV and a 0.72 for email carry the same strength of evidence, and a person's channels can be read side by side. The line sits at 0.25, so a score can fall at most 0.25 below it and can rise as much as 0.75 above it.
Averages hide the two groups a planner needs.
An average says a channel is good or poor for an audience as a whole. Inside a channel that is good on average, some people score below 0.25, and the money spent reaching them there buys little. Inside a channel that is poor on average, some people score well above the line and can be reached there at that channel's price. Scoring each person shows both groups. An average shows neither.
Two channels a point apart are tied.
When two channels score within a point or two of each other for a person, they are tied. Run both and let price decide. A ranking puts one first and hides the tie, and the tie is often the most useful fact on the file, because it lets the buyer take the cheaper channel and give up nothing the scores can see.
Pricing the spend below the line.
The spend a plan can move off a channel is the number of people it pays to reach there, times the cost per contact, times the share of those people who score below 0.25 on that channel.
× cost per contact
× share scoring below 0.25 on that channel
Only the share comes from the scored file. The file size and the cost are the client's own, which makes the result the client's number. The share is also the figure a client can check against its own delivery report. The spend does not disappear: it moves to the channels where the same people score above the line, and the shift shows in the next report.
Score each person, then draw the line.
Score each person on each channel, draw the line at 0.25, and take the people below it out of that channel's buy. Channel scores travel in the super audience file or append to a client's own list. To see them on your audience, book a Data Discovery.
Book a Data Discovery →Continue to 08 · Your Data, Your Deployment →