Revisiting the measurement of mail in econometric models

07 Mar 2019  |  Louise Cook 
Revisiting the measurement of mail in econometric models

Sponsor content: Louise Cook explains what JICMAIL means for the industry from an econometric point of view

The race to understand how digital channels impact business KPIs has left little bandwidth for pushing econometric frontiers elsewhere. But it is time to look again at mail, both addressed (Direct Mail) and unaddressed (Door drop).

Why now? The launch of JICMAIL finally gives mail its own media planning currency. With that comes fresh evidence, now based on over 80,000 mail items, showing how consumers interact with mail and the type and timing of commercial actions it prompts.

The emergence of programmatic Direct Mail means mail both growing and being used in new ways, presenting new types of consumer responses to unpick. This is something we should now make a priority.£1.7bn is invested annually in Mail with more spent only in TV and internet.

From an econometric perspective, mail has often been the medium whose results provoke most dissension. Mail users often measure mail initially via a trackable URL, code or phone number. This can be a bit like digital’s “last click” attribution.

The Private Life of Mail (2014) demonstrated Mail acting in ways similar to more traditional offline media, eliciting both response and emotion, making consumers feel valued and making them remember longer term the brand which mailed them. (It was this work which led to JICMAIL being developed.) Econometrics tries to capture these broader Mail impacts, but its results are often regarded as less factual than URL based evidence.

Although the JICMAIL database’s primary purpose is to provide planning metrics for targeting mail, it will benefit econometricians too. By providing a common campaign language, the strategic intentions behind a mail campaign can be more effectively shared.

This should help us produce econometric results more focussed on planning outcomes and so more useful. And by providing an estimate of audience impressions, JICMAIL enables better comparison of Mail’s impact with other media. Additionally, the behaviour patterns in the JICMAIL data will help to reconcile, econometric results with other measurements of Mail’s impact.

Sense-checking econometric mail results can be difficult, particularly when different sectors use mail differently and when item reach varies by sector - magazine/newspaper mail achieves over 25% higher reach per item than either Travel or Financial Direct Mail. The JICMAIL data thus provides a useful cross-sector guide.

Applying JICMAIL data

The JICMAIL data will add value at every modelling phase by influencing the way we incorporate mail into models. It should help with the data request, with expressing mail’s effects mathematically, assessing whether outputs are valid and informing how we report and benchmark results.

It will give us far more options for legitimately exploring the way mail works, so should discourage us from discarding some (possibly counter-intuitive) mail impacts early on in the modelling process.

Ultimately, econometric results which all parties acknowledge are representative, will enable seamless optimisation across all media, including mail.

Mail evaluation checklist

  1. Briefing - ask the mail planner to brief you on campaign strategy and the key insights drawn from the JICMAIL database.
  2. Selecting input data – based on campaign strategy and how mail has been shown to work in the sector being modelled, consider how mail data should be incorporated – as a whole, by message type, by customer type
  3. Expressing mail effects – formulate your hypotheses as to mail’s effect and how it should be included in the models based on JICMAIL information showing:
    a) how quickly customers react initially
    b) how drawn out effects are expected to be over the next few weeks
    c) where effects are concentrated
    and then on your
    d) ability to report on the specifics of the mail strategy deployed
  4. Validating models – do measured effects reflect relevant JICMAIL patterns? If not, is there a good reason why not? Are models of KPIs most likely to be affected by mail showing those effects or are effects being masked by other factors? Can you bridge the outputs from econometrics and URL-based reporting?
  5. Reporting – how do econometric results reflect strategy? What can you offer which endorses that strategy or points the way for future development (including the role of mail in the mix)?


Louise Cook is Managing Director, Holmes & Cook

This is the first of “Planning with JICMAIL” – a series of articles and guides by industry experts and practitioners that seek to unlock the value of JICMAIL data for advertisers throughout all stages of the planning cycle.

Advertisers can access JICMAIL data for a free subscription period. If you’d like to see the value of JICMAIL data or attend our event, please contact tara@jicmail.org.uk or go to www.jicmail.org.uk.

Leave a comment

Thank you for your comment - a copy has now been sent to the Mediatel Newsline team who will review it shortly. Please note that the editor may edit your comment before publication.