Customer remarketing for ecommerce websites

Basket and checkout abandonment is a natural part of ecommerce life. The average abandonment rate in the UK is between 60% and 70% (depending on where you source your stats – clients I’ve worked with typically range between 50% and 65%).

Why is this? Is it because ecommerce checkouts provide a poor UX?

The answer is nuanced. For the sake of brevity, let me break it down into three common causes:

Poor UX

If users encounter unexpected (or unnecessary) friction during the purchase process, they’re more likely to leave. For example, forms that aren’t optimised for use on mobile devices, such as card payment fields not defaulting to a numerical keypad.

Technical glitches

Some errors are out of your control, for example connectivity issues with a local exchange or intermittent problems with mobile connections or WiFi. If a user is struggling to get the site to respond, they’re more likely to abandon.

Customers not ready to buy

Research is expected before committing to an online purchase. This can involve creating a basket on multiple sites and then comparing the full costs, or creating a basket that you want to edit and refine at a later date before you’re ready to buy. These are genuine reasons why a customer might create and then abandon a basket before completing the checkout.

The role of remarketing

Remarketing is a proven technique for catching users at the point of abandonment and giving them reasons to convert either in the same session, or by prompting them to return to the website via targeted, personalised marketing campaigns.

This blog looks at the key elements of a remarketing program and offers tips and advice. Follow, adapt or ignore as you see fit!

Please do add your experience and insights via the comments thread at the end.

1. Exit intent prompts

This is your starter for 10. At the most basic level, it’s an overlay that is displayed to users who have an active basket when they close their browser session. To see the overlay, users have to accept the browser prompt, so you can’t guarantee that every user will see your content.

The most common exit overlay is a simple reminder that there are items in the basket with a value proposition to encourage conversion e.g. reminding users of free delivery and returns. The example below is from business software company, Sage.


However, relying on overlays for abandoned baskets misses a much bigger set of users; those who exit the site before creating a basket. There are two options here:

1. Trigger overlay prompts when a user exits the site earlier in the funnel e.g. on a product page

The advantage is that you can tailor the content to the product being viewed, for example showcasing customer reviews for that product.

2. Use exit intent predictive analysis to automate prompts before users exit

The aim of exit intent analysis is to learn from user browsing behaviour to understand what actions typically lead to an exit. This data is then used to predict when a new user is likely to exit and proactively trigger a prompt before they exit, which avoids the need for the user to accept the browser prompt to see the message.

It’s not an exact science but it does increase the user pool for remarketing on-site and you will learn over time what works and what doesn’t i.e. if prompts don’t nudge conversion and have a detrimental impact on metrics like bounce rate, reconsider. There are specialist tools available including Yieldify and Rooster.

Key pointers for exit overlays:

  • Minimise content – users are leaving, don’t expect them to read War and Peace
  • Include a prominent CTA
  • Make it visually appealing – don’t let it looks like a standard browser message.

2. Remarketing emails

This relies on having the user’s email address, either from a new user entering it during the checkout or an exiting customer signing-in before exiting (which also includes soft log-in state – ‘remember me’).

If you’re running exit overlay and emails, be sure to remove people who respond to exit overlays from your email remarketing pool.

Some pointers on what I’ve seen work well:

  • Multi-stage campaign (3 emails seems to be optimal) – the cumulative conversion is greater than that of a single email
  • The first email is customer service focused and sent within 2hrs of the exit – make sure they didn’t have problems ordering and offer support if they did
  • The second email uses more persuasion (eg. promoting key selling points) and is sent after 24hrs
  • The third email is more sales focused (often promoting an offer or emphasising key selling points like warranties) and is sent after 48hrs
  • Emails must use responsive templates to ensure they render well on mobile
  • All links must have campaign tracking codes
  • All emails need to have a clear CTA at the top (eg. ‘Complete my order’)
  • The first email should show the basket contents
  • Don’t include a discount incentive in the first email – you could sacrifice margin unnecessarily (although I’ve seen marketers ignore this to drive short-term revenue).

3. Remarketing display ads

Display ads can be scheduled on any ad network, including social media channels like Facebook. With display, you get the volume as you don’t need an email address and you don’t need users to accept browser prompts. You use a cookie that will recognise the user’s browser and device, and serve ads as they browse other websites (a highly simplified explanation).

Dynamic remarketing with AdWords allows custom variables and custom feeds to tailor ads to suit your product range and audience. You can also exclude people from remarketing lists to help minimise cost wastage, for example a subscription service could exclude existing subscribers from brand campaigns (to avoid paying for clicks from people who are already signed up).

With remarketing lists you can upload customer emails into platforms like Google and Facebook and then create remarketing campaigns targeted to those users. This helps you segment and personalise the remarketing ads. For example, you can target loyalty scheme members who abandon baskets with ads that show the loyalty points/rewards they can earn by completing the purchase.

The example below is from Wayfair, who are merrily stalking me around the web and reminding me that I can buy from them still if I want.


If you’re using Google Analytics, you can turn on Remarketing in the admin panel and set up lists based on a wide range of metrics, which can then be imported into AdWords to extend your remarketing audience.

4. Remarketing driving offline sales funnels

In B2B ecommerce, online isn’t just a direct sales channel, for some brands it’s also a key driver in the sales funnel, generating leads for the sales team.

Let’s use the example of finance and accountancy software. With a purchase that a) can costs £thousands and b) impacts the entire business, the decision making cycle is typically much longer than a simple retail purchase and in large organisations involves multiple people (influencers, budget holders, decision makers etc.).

So, if a user adds a software bundle to the basket but then abandons, it’s important to test promoting the offline support channel as a response mechanism, rather than assuming remarketing focused on online conversion will yield the best results.

A few pointers:

  • Set-up a dedicated trackable phone number for remarketing campaigns (ideally use a call tracking tool like ResponseTap so that you can generate unique numbers for each lead to allow more granular analysis)
  • Promote the key benefits of speaking to a dedicated customer support team (and/or promote Live Chat help)
  • Brief the sales support team and ensure they’re ready to handle the extra calls (and know where they’re coming from)
  • Test showing a contact number vs. asking users to submit their phone number for a free callback (and make sure you use validation on the phone field so you capture valid numbers!)
  • Ensure your CRM system can match offline sales back to online leads – can be as simple as passing a URL parameter through to the CRM.

Please note I’m not saying B2C ecommerce can’t benefit from using remarketing to drive offline sales, just that in my experience this is a more common use case for B2B. I can think of some interesting options for B2C luxury retail in terms of promoting in-store personal shopper services for certain basket combinations (though I’ve not tested this yet).

5. Planning and measurement

As with any digital marketing, it’s crucial to have a clear plan for remarketing, with all stakeholders briefed and working towards the same goals.

Things to think about include:

  • Build a launch timeline that maps key project steps and deliverables (with owners) to get you live
  • Define which products/services are you remarketing for
  • Decide if you’re building out the remarketing capability in-house or through a 3rd party provider like VE Interactive
  • Data – what data do you have on current abandonment (not just site wide but also down to product level)?
  • Financial model – build out a revenue and ROI forecast based on this data, factoring in the costs of the project (ensure key inputs like traffic volumes are editable so you can flex the model based on different thresholds)
  • Content – have a clear plan for your launch campaign content (brand assets, copy, images, CTAs etc.); start simple, improve through testing
  • Create a test plan – a matrix of tests you will run after launch to improve open/click/conversion rates
  • Run hold-out tests to prove revenue incrementality – exclude a subset of the potential audience from the campaigns for comparison
  • Set-up reporting – ensure web analytics and reporting tools are set-up to capture the campaign data and provide clear reports.

6. Campaign controls

You need to be careful with remarketing that you don’t bombard users with annoying campaigns. Sadly this is common and although it can’t be 100% avoided, there are actions you can take to minimise the risk of remarketing stalking.

These include:

  • Setting frequency caps for display remarketing to ensure users aren’t constantly shown ads for an indefinite time period
  • Also use a frequency cap for exit overlays to control how often an individual users sees the prompt (eg. only once per day for a maximum of X days)
  • Use blacklists to exclude users from remarketing e.g. use CRM data for offline purchases.

Comments and questions

So what do you think are the key requirements for doing remarketing effectively? What are the risks and how can you mitigate them?

Please drop by and share your comments, questions and experience.