How to Decode the Offline/Online Match Process

joannaoconnelrevised“Marketer’s Note” is a weekly column informing marketers about the rapidly evolving, digital marketing technology ecosystem. It is written by Joanna O’Connell, Director of Research, AdExchanger Research.  

Digital marketers are sitting on a gold mine – their own organization’s first party data– and they are finally starting to tap into it. Chances are, millions of dollars have likely been spent in the building, cleansing, segmentation, and activation of this precious marketing resource in offline and other CRM channels. As organizational siloes continue to crumble, we see more and more companies extending the use of their traditional first party data sets into digital environments, whether for suppression of messaging to existing customers, reactivation of lapsed customers or the like (there are lots of cool applications of this data!)

But, through my regular conversations with marketers, I get the sense that many find the actual process of translating their data – which often includes personally identifiable information, or PII – into usable, anonymized digital data murky and hard to understand, let alone explain to colleagues. To that end, I’d like to share some questions I’d ask potential match partners (such as LiveRamp) if I were considering them for such an endeavor:

  • What is the step-by-step process for performing the match?
  • Where does your offline data come from (Is it proprietary? Licensed from 3rd parties?)? What form does it take? How do you verify its validity and quality?
  • Where does your online data come from? i.e. How do you build and maintain a matchable cookie pool? Do you have a stable suite of partners with whom you have direct relationships? How, if at all, do you validate cookie stability and quality?
  • What kind of scale can I expect, and at what point in the process?
    • What kind of match rates do you typically see between your data set and client data sets?
    • What percentage of matched cookies are you typically able to reach?
    • Is there some kind of ramp period?
  • Do you match iteratively? What’s that process? How frequently?
  • What resources should I expect to make available from my side?
  • What contractual and legal processes – if any – will we need to go through?

As always, the above list is not intended to be exhaustive, but it should prove a helpful place to start in your efforts to decode the offline/online match process!

Thoughts, comments, send them my way!


Follow Joanna O’Connell (@joannaoconnell ) and AdExchanger (@adexchanger) on Twitter. 

Enjoying this content?

Sign up to be an AdExchanger Member today and get unlimited access to articles like this, plus proprietary data and research, conference discounts, on-demand access to event content, and more!

Join Today!


    • (continued)….you can match CRM data via the MD5 Hash of an email address, a process completely independent of cookies. A la, Facebook Custom Audience, Twitter Tailored Audiences & LiveIntent LiveAudience…

  1. We hear a lot of confusion in the marketplace about offline/online matching too. We don’t offer this as a stand-alone service, but we do a fair amount of this kind of work for our clients when they want us to market to an existing customer file… A few things to also consider:

    1) Pricing model – some providers charge based upon the number of records on the input list, not on the total number of matches. Others charge only when a match is found. So it’s really important to understand exactly how the billing process works. Is this a one-time charge or monthly? Does it cost extra to have the on-boarded audience distributed to more than one destination point?

    2) Matching rates. This one also has huge pitfalls if you’re not careful how you ask the questions. Providers have the ability to match at different levels of granularity (individual, household, or even neighborhood zip+4). Obviously, you make more matches if the criteria for the match is looser. Neighborhood is looser than household is looser than individual. Once an entity is ‘matched’ it is possible to associate multiple browsers and/or devices with the same entity. So just be careful what you ask for. Here are 2 real-world examples:

    100 name, addresses, and emails provided as input. Specify a neighborhood match.
    Found 300 households within the specified neighborhoods
    Found 546 individuals within those households
    Found 1556 cookies associated within those individuals
    Depending upon how you look at it your “match rate” is 1556%

    100 emails provided as input. Specify an individual match
    Found 23 emails that match
    Found 94 cookies associated with those emails
    Now my real match rate is 23%.

    NOTE: We haven’t even started to factor in cross-device mapping into this calculation!!

    Neither matching scenario is wrong – it just depends upon the use case. Your advice to make sure you understand the step-by-step process probably covers these 2 use cases, but these are 2 particular points that might not be as obvious to your readers.

    • Joanna

      Thanks so much for weighing in, both Scott and David!

  2. I know this post is a bit dated at this point but did you receive an answer from companies like LiveRamp?

    • Joanna

      Hey there Nick – not really, just what you see posted here!