Match rates come up in every evaluation of data onboarding service. Though there is no industry standard in calculating match rates, this number is critical to understand because it determines how much data can be activated and put into use in online marketing platforms.
Here are the two key points you should focus on when it comes to match rates:
1. How do you define your match rate?
You may encounter more than one definition of match rate. Some companies use a single point match where it counts all separate devices as one match each – for example, an iPhone, work computer and home computer. Others use a multipoint match where one match is equivalent to matching all the three abovementioned devices to an individual.
Another key difference to note is between match rate and sync rate. The sync rate, also known as ‘cookie reach’, is the rate at which you can connect an individual’s identity to devices, from one to often several.
For an example, if you sent a million CRM records to your onboarding partner, they may say they have reached a million devices. You probably think that this equates to a 100 per cent match. However, in reality, the true match rate could just be 20 per cent. But by the time the onboarder has pushed that identity to a data management platform, they have on average five devices associated with that identity (5 devices x (1 million x 20 per cent) = 1 million devices, not individuals).
This ‘sync rate’ gives you a feel for the number of ‘bites of the cherry’ you may have for targeting an individual, but it is not the same as the number of individuals precisely matched, which we believe this is the true ‘match rate.’
In a nutshell, the match rate is, or at least should be, the percentage of individual matches made between the onboarded file and the consumer in the digital world. Of course, you can always define your own match rate if it makes sense for you and your consumer.
2. What is the level of accuracy of your onboarding partner?
It is important to understand the precision level that a data onboarder can truly support. It is also wise to be wary of partners who make accuracy claims with IP matching – unless you are doing broad geo-targeting or targeting employees at large businesses with static IP addresses.
And why? Because deterministic match data is difficult and costly to acquire at scale. Many data onboarding services use probabilistic models that rely on IP address matching. Here, matches between an IP address and a cookie can be easily collected any time an ad is served or a browser page loads.
Probabilistic Matching uses statistical algorithms to quantify the likelihood that the person being targeted is the same as the individual being onboarded. Probabilistic Matching may use postcode level data or IP addresses, device behaviour and the like to predict the likelihood they have the right individual.
Deterministic Matching uses a wider and richer set of identity resolution techniques and data, typically using authenticated user login data, to resolve that identity to the individual level. It is not probably that person; it is determined to be that person.
However, for these matches to be useful for data onboarding, the IP address must be mapped to personally identifiable information (PII). This is done by geolocating the IP address and matching to a street address – a process that is very inaccurate.
As you can imagine, deterministic matching is more desirable because it is more accurate and so delivers better results through greater relevance. However, there is a place for probabilistic matching so long as you understand the differences.
The definition of a match rate varies by company, along with how to accurately determine the match rate. Other than understanding what a match rate is and why it matters, here are five questions every marketer should be asking their data onboarding partner about match rates.