Hit rate, and the numbers worth measuring instead
Every provider is asked for a hit rate and every honest answer is "it depends on the county", which satisfies nobody. Here is what to track instead.
Why hit rate under-informs
A hit rate is a ratio of matched records to submitted ones, and both halves are defined by whoever is quoting it. Count landlines and it goes up. Count entities in the denominator and it goes down. Compare across counties and it moves for reasons that have nothing to do with the provider.
It also says nothing about whether the matched numbers are current, which is the property you actually care about.
We do not publish a single figure for this, because a single figure would be true of some counties and false of others, and an average presented as a promise is a fabrication with a decimal point on it.
Cost per usable row
Take the total you paid and divide it by the rows you would genuinely put into a dialer — mobiles, people rather than entities, in the market you meant to work.
This collapses every pricing trick into one number and is the only figure that survives contact with a campaign.
On our files it is close to the sticker rate by construction: unmatched rows are credited back, landlines and VoIP are dropped before billing, and entities never enter the file.
Contact rate, conversation rate, and cost per conversation
Contact rate is the share of rows where a live human answered. This is where data quality actually shows up, and it is the first number that will tell you a file was old.
Conversation rate is the share of contacts that became a real discussion rather than an immediate no. This measures your script and your market far more than it measures the list.
Cost per conversation is the two combined, and it is the number to compare between providers and between list types. It is also the number that will tell you a cheap file was expensive.
Questions people ask
What hit rate should I expect?
Enough varies by county that any number quoted as a national figure is a marketing statement rather than a measurement. Compare providers on the same county in the same week, and judge on cost per usable row.
How many records do I need for a fair test?
Enough to call all the way through, which for most people is a few hundred rather than a few thousand. A test you do not finish measures your stamina, not the list.
Is a low contact rate the list or my calling?
Look at when you are calling before blaming the data. Time of day moves contact rate more than most list quality differences do.
Related
Pick a county and a list type, or send your own addresses. Mobile numbers only, litigator and DNC scrubbed, and anything the county does not hold is credited back.