Using propensity scores for email targeting
Email marketing. These days, everyone does it. But only a handful do it brilliantly.
The reasons for its popularity are clear: low cost to produce and execute, fast to deploy and totally measurable. The methods behind brilliant execution are no less clear – it’s just that few people know about them.
Before we share the secrets of brilliant email marketing, let’s take a step back and look at the massive paradigm shift that has taken place in the marketing world.
Email rewrites marketing rules
Sure, some of the basic tenets of marketing still apply in the online world. Know your customer; uphold your brand’s reputation; and keep the communications coming.
But email marketing changes the way we execute campaigns. Previously, a marketing team would come up with a concept for a campaign, conduct market research, refine the concept and then launch it to their entire customer database.
Now you start with a concept for an email campaign, send it out to a test audience, adjust it based on response rates, send it out to the next batch, adjust it again, send it again, and so on.
Instead of sending one campaign to your entire database, you are in a continual cycle of test-and-improve. The concept keeps evolving, and the response rate keeps improving.
It takes a while to get such an email program right, and requires an ‘innovate and experiment’ mentality. Most marketers have these traits in spades anyway – you just need to shift towards an ‘always on’ attitude where the campaign is never over.
With this shift in attitude, you’re now ready to learn some of the secrets of brilliant email marketing.
Propensity scores boost response rates
If you know your customers’ likelihood – or propensity – to engage with an email, you’ll achieve greater success with email marketing programs and greatly facilitate this test-and-improve cycle.
Propensity modelling is a proven technique, used in everything from medical research to direct marketing. But it’s been neglected in email marketing… until now.
At Bienalto, we build propensity models for email marketing that takes all aspects of email communication – such as campaigns, event invitations, and newsletters – and then collects data on individuals’ behaviours and profile.
The models give us a percentage figure that represents their likelihood to respond to the next email, which in turn gives us a stable dimension for selecting who to target – and who to cull – in the next campaign.
Why is it important to cull the people who won’t respond to an email, when it’s so cheap to send the campaign to everyone on your database?
Think about it. Every time you send an email to a customer, you are testing the relationship. If you send them something they don’t want, it’s more likely they’ll opt out. And once they’ve opted out, it’s pretty unlikely that they’ll opt in again down the track.
Use propensity scores to get a better idea of what your customers want, and only send emails to those you know will welcome them and are likely to respond. It takes segmentation to the next level, by continually monitoring a person’s behaviour and profile, and adjusting campaigns to match this behaviour.
Setting up a propensity model requires the marketer to practice closed-loop marketing techniques. And, unfortunately, many organisations lag in this area.
Whether it’s a case of marketing and sales not ‘talking’ to each other about their email campaigns; or failing to clean up databases by removing bounced email addresses; or failing to capture click-throughs and leads from a campaign; or simply not setting your CRM up in the right way – these are classic symptoms of a hole in the loop.
Close the loop to extract meaningful data from your campaigns, so you can:
- Select, segment and compare profile characteristics
- Create propensity scores
- Conduct ‘what-if’ financial scenarios based on response models
- Target only those people more likely to take the action you seek.
Propensity modelling is an ongoing process, which needs to be updated on a regular basis. We also find that each model need to be custom-built specifically for each client, reflecting the unique marketing needs and goals of the business.
Re-marketing opportunities abound
Email marketing also presents incredible re-marketing opportunities, where you can supplement an initial campaign by sending an additional piece of communication based on who clicked what in the first email.
Or you may re-send unopened emails with a different subject line or without graphics, after a suitable period of time has lapsed. It’s about getting the most mileage out of your campaign and casting a wider net to capture potential customers.
And it works. On average, our re-marketing strategy delivers a 25% increase in the overall response rate of email campaigns for our clients.
In addition to re-marketing, you should also consider the number of pathways to a desired action. You may have one flagship email communication strategy – for example, a monthly newsletter. This email acts as a secondary device for promoting other campaigns. You can then pick up the people who didn’t respond in the flagship email, and target them with another email.
Email marketing presents endless opportunities for connecting with your customers, building loyalty and generating meaningful dialogue when it counts.
Learn more about Bienalto’s marketing programs.




[...] review: Akin Arikan’s Multichannel MarketingUsing propensity scores for email targeting’Always beta’ sites are invariably betterMultivariate testing: what, how, whyUser experience at the [...]
[...] Q and A with Paul O’Connor, Marketing Director, Sun MicrosystemsAMP website wins Rainmaker awardsPreparing for a site re-design: don’t forget analysisBook review: Akin Arikan’s Multichannel MarketingUsing propensity scores for email targeting [...]