← All playbooks·Media planning

What you need first

• Program's rating among the target audience for the measurement period, from panel data (AGB Nielsen, GfK, Médiamétrie or equivalent)
• Program's rating among total population/universe for the same period, same panel source
• Target audience universe count (population estimate for the demo), from the panel's establishment survey
• Total population universe count, from the same establishment survey
• Underlying sample size for the target cell in that period, from the panel report, to judge reliability
• Shortlist of competing programs on matching daypart and universe base, from the buying grid

The procedure

1. Pull the program's target-audience rating for the period from panel data; this produces the target rating (%)
2. Pull the program's total-population rating for the same period from the same source; this produces the total rating (%)
3. Divide the target rating by the total rating and multiply by 100; this produces the affinity index
4. Check the target cell's underlying sample size against the panel's minimum reporting threshold; this flags an unstable index before it goes further
5. Compare the index to 100; above means the program over-indexes for the target versus the general population, below means it under-indexes
6. Rank the index against the other shortlisted programs on the same universe base; this produces the relative fit order for the buy
7. Read the affinity index alongside the program's absolute rating or GRPs; this produces the final call on whether the over-index is worth the spend

Worked through with numbers

Target: women 25-54, universe 3,200,000. Program target viewers 224,000, so target rating = 224,000 / 3,200,000 x 100 = 7.0%. Total population universe 14,500,000, program total viewers 580,000, so total rating = 580,000 / 14,500,000 x 100 = 4.0%. Affinity index = 7.0 / 4.0 x 100 = 175. A reading of 175 means the program delivers women 25-54 at 75% above its delivery to the general population, a strong over-index worth checking against its absolute rating before committing spend.

Where it goes wrong

• High affinity on a program with a tiny audience gets treated as a strong buy; correct this by reading the index next to absolute rating or GRPs, not on its own
• Affinity indices built on different universe bases (national panel vs regional panel) get compared as if equivalent; correct this by confirming both programs use the same universe before ranking them
• A single week's spike gets reported as the program's affinity; correct this by averaging across several measurement weeks before quoting a figure
• A rating built on a handful of target respondents gets treated as solid; correct this by checking the target cell's sample size against the panel's minimum threshold before trusting the index

How to know it is right

Terms used

• Affinity Index
• Rating
• Universe

← All playbooks Calculated in TV Planner →