Overview

To understand a baseline match rate (and thus, the addressability of those customers) for YOUR customer data within the DSPs & AdTech systems you use, it is recommended to run a few experiments using Wondaris. This is also a good way of understanding how nimble you can be now that you can use Wondaris.

These experiments will allow you and your team to understand how to think about your audiences, including filters, size & included columns to get the most out of your first party data, as well as provide you with guidance into improving your customer data.

Background

To utilise first party data in advertising systems, various identifiers are synced into the Demand Side Platform (DSP) or AdTech system for matching. If these customer identifiers are not present and able to be matched inside the DSP, then the customer is not  included in the targeting within that DSP.

Match rates vary wildly across DSPs and identifiers - some DSPs can handle more identifiers to help improve the match rates (providing more certainty of matches) whilst others can only handle a few. Sometimes, adding more identifiers can negatively impact the match rates.

The most common identifier used is hashed email, closely followed by hashed phone number (although this is often harder to ensure the formatting is correct globally).

The general rule of thumb is to try to get at least 1000+ matches in your audiences within the DSPs to ensure a decent audience size - noting that a lot of DSPs will not allow you to target below this number. You may, however be able to use lower than these numbers for things like lookalike audiences as the DSPs can expand the lookalike criteria to lessen any impact specificity in your audience has (ie: the DSP can be looser with the look-a-likes to get broader audiences from the seed audience).

Analysis Strategies

To help analyse and possibly improve the match rates, there are a number of strategies that you can take to more deeply understand what, within your data impacts the match rates within the various DSPs.

What you will want to understand is the following:

• The general match rates you get with your data in the various destinations

• Which segments in your data impact the match rates in the various destinations

• Which matching keys (ie: columns) in your data impact the mate rates

To do this we propose determining a series of experiments - these experiments will focus on defining & sending “match test” audiences to the DSPs to see what the differences in match rates are. The outcome of these matches will then be used to form further analysis strategies or to document how to think about creating audiences with first party data. The analysis will also form the basis of decision making in improving the match rates through data  collection & hygiene.

Match Rate Analysis Process

Below are the high level steps to follow when doing the analysis work (along with the responsible parties)

1. Define the audience segments to test - Client / Implementation partner or Wondaris CSM

2. Define the destinations to test in - Client / Implementation partner or Wondaris CSM

3. Create the Audiences in Wondaris - Client / Implementation partner or Wondaris CSM

4. Create the destinations & lists in Wondaris - Client

5. Activate the audience into the destinations - Client / Implementation partner or Wondaris CSM

6. In 24h - 48h get the match rates from the destination systems & document the results into your business' marketing knowledge base (be it a spreadsheet, a knowledgebase or some other documentation system) - Client

7. Analyse results & determine next steps

   1. Document findings as far as general match rates are concerned

   2. Plan other segmentation to test

   3. Document segmentation strategies for future audiences

   4. Highlight what could be done to improve data collection / hygiene

The output of this work will be a better understanding of the match rates and how your first party data impacts those match rates - this output can take the form of a playbook for using 1st party data as well as tactical plans to improve data or further test match rates, and should be shared amongst your business' marketing team members (as well as others who may be responsible for data collection & hygiene).

What to think about

At each of the steps above these are the things that should be thought about when planning this work:

Defining the audience segments to test

• How your data is currently segmented now and how you will think about using your data for audiences in the future (to inform future activities)?

• Some high level segmentation to start with could be:

  • All customer data (mandatory - make sure you include this)

  • Segment based (what segmentation do you have now in your customers - ie: customer type, age, etc)

  • Region Based (do you segment users based on global regions or even national regions)

  • Value Based (what value metrics do you use to segment customers - ie: number of purchases, lifetime value, yearly spend, average spend, etc)

Define the destinations to test in

Test in as many destinations as you want to, but try to include all those you will use in the future to give you a baseline understanding of for the match rates.

Audiences & Destination lists

Make sure the audiences in Wondaris & the customer match lists in the DSPs names are clearly labelled with an appropriately clear label indicating their purpose.

We recommend something like “Match Rate Testing - { AUDIENCE }” which clearly indicates it is a test audience and should not be used for advertising.

Activating the audience/s

Only do a single send - this can be achieved in Wondaris by an instant activation schedule or a single future scheduled send (ie: non-repeating schedule).

Improvement Strategies

To help improve the match rates a number of strategies that focus on data collection & hygiene can be employed. These should be determined once deep analysis has been completed, but a high level approach could be as follows:

Data Collection

Improve overall data collected:

• Ask for backup emails (secondary email for account recovery)

• Ask for backup phone numbers (also for account recovery)

Data Hygiene

Improve the hygiene of the data:

• Asking customers to confirm their details in your platform / app & on the website

• Running an outbound call centre campaign to double check customer details & update in your CRM

• Send emails / SMS out to ask customers to keep their details up to date

• Ensure emails are correct & hashed appropriately