Love is… a multi-parameter engagement algorithm

9 minute read

Love is in the air. So now we know each other it’s only right that we discuss getting engaged. Engagement (not least in email marketing terms) is a tricky concept. Traditional email performance metrics like open and click-through rate give useful campaign-based feedback, but to get anything more than a basic understanding of engagement you’ll need to measure a number of different factors and over a prolonged period of time.

The Coffee Break Guide to Engagement

With the right approach, and the right tools, engagement can not only be measured, but managed and actively exploited. We’ve put everything you need to know about email engagement into our handy guide. It explains the concepts and application of engagement profiling in simple, easy to follow steps.

Guide to engagement

  • What does engagement look like?
  • How can it be measured?
  • What insight can we gain from engagement profiling?
  • How can engagement be used to create precision targeting?
  • What are the benefits?

Grab a coffee and download your free copy from our website. Meanwhile… here are some extracts.

What is engagement?

For the moment let’s start by characterising engagement into 3 broad categories (when we come to define engagement more specifically we’ll expand these to 5 distinct engagement bands). Each of these 3 categories can be broadly classified by observing the typical behaviour of subscribers, both in relation to their levels of campaign interaction and their extended social relationships.

Highly engaged

  • Regular opens of your email campaigns.
  • Frequent click-through to additional linked content.
  • Recent completion of post-campaign goals and transactions.
  • Share content socially. A brand champion to peers and colleagues.


  • Less frequent, occasional opens. Inconsistent interaction.
  • Irregular click-through history to additional content.
  • Low to average post-campaign completion actions taken.
  • Generally non-committal regarding recommendation to others.


  • Habitual non-opener or deleter of your campaigns.
  • No recent click-through history to additional online content.
  • No recent purchases or online content downloads.
  • No current interest. Potential unsubscriber and brand detractor

Precision targeting – Audiences and Dimensions

Engagement profiling is actually one of many ways you can add precision targeting to your email campaigns. Precision targeting uses the concepts of ‘audiences’ and ‘dimensions’ to create dynamic data segmentation and targeting.

Audiences is a way of defining segmented subscriber groups according to specific profile characteristics. Dimensions are the various ways in which that profiling information can be defined and gathered.

dimensions (600px)

In there are 7 profiling dimensions that you can use to define a targeted audience.

  • Date – filter by initial subscription date.
  • Geo-location – the physical whereabouts of your subscribers.
  • Lists – include or exclude from multiple stored lists.
  • Profiles – any data captured in a subscriber profile.
  • Frequency – according to recent communication history.
  • Domain – group by company domain names.
  • Engagement – by historical levels of campaign interaction.

How can engagement be measured?

The problem with engagement is that it’s difficult to measure in an objective way. It depends on many different factors, each of which is subjective to a lesser or greater extent. It applies both to individual subscribers and to audience groups, and it’s not static – engagement changes over time.

At we use a multi-parameter algorithm which monitors a range of engagement characteristics. It takes into account a broad view of engagement behaviour, including the frequency and depth of interaction with each campaign. It’s fully automatic – the engagement algorithm continually runs in the background collecting interaction data from each executed campaign. The collected data is processed to create an engagement profile for each subscriber – we call it their engagement rating. Based on their history of interaction each subscriber is attributed with an individual engagement star rating.

  • 5 star – highly engaged
  • 4 star – mildly engaged
  • 3 star – passive
  • 2 star – mildly disengaged
  • 1 star – disengaged

New subscribers are initially assigned a mid-range, 3 star classification. There is also a ‘null’ classification for subscribers for which there is no qualifying data available. You’ll see your subscriber engagement star ratings when you review your lists in the ‘collect’ section of the platform.

Engagement star ratings

Here’s a little more detail on how the engagement rating algorithm works.

  • In order to accurately reflect a range of observed characteristics the engagement rating bands are based on both theoretical and matched empirical data.
  • Starring bands are non-linear between maximum and minimum values – again to better reflect observed situations.
  • Each campaign interaction (open, click, share etc.) earns a weighted increment depending on the action taken.
  • Opt-in is considered as a key engagement indicator.
  • Multiple interactions with a single campaign gain additional engagement increments.
  • Passive behaviour (non-interaction) over a period of time is considered as a negative characteristic. Non-engagers progressively decay in their star rating over time.
  • Habitual non-engagers decay proportionately more rapidly.

The advantages

Engagement profiling removes the need for guesswork or gut-feel and provides a simple, objective measure of complex behavioural patterns.

  • It’s a quantitative measure of an abstract concept – the 5 star rating provides a simple interpretation of the data.
  • The multi-parameter algorithm provides a holistic view – it takes account of a wide range of interaction behaviour.
  • It’s a fully automated process – as a user you don’t need to collect any data or manually take action.
  • Engagement behaviour is monitored and subscriber data is collected over an extended period of time.
  • The engagement star rating is dynamic – results are continuously updated as new campaigns are executed.

Engagement profiling can be used to target VIP rewards, exclusive offers and loyalty scheme awards to your 5-star subscribers. Similarly you can use your profiling intelligence to re-engage with subscribers who are falling away from your brand. Entice them back with special offers and ‘welcome back’ discounts. By definition engagement is likely to be low but some will return and others can be eliminated from future campaigns.

VIP campaign

Engagement – in a nutshell

  • Engagement profiling is a precision marketing technique designed to deliver relevant campaigns based on subscribers’ engagement history.
  • Engagement is one of seven profiling dimensions that can be used to create audience segmentation and targeting.
  • It uses a multi-parameter algorithm to capture and quantify both the quantity and quality of subscriber engagement over time.
  • Every subscriber is assigned an engagement star rating, from 1 star (disengaged) to 5 (highly engaged).
  • Engagement profile data is captured automatically and subscriber ratings are updated continuously as campaigns are executed.
  • Profile ratings can be used to target campaigns to specific engagement groups, for example with VIP rewards or re-engagement campaigns.

If this has set you thinking, there’s more information in the full downloadable guideIt’s free.