Calculating Conversions With Cookies
Most commonly, this is set to 4 weeks / 1 month and attributes conversions to the last email clicked, which could actually be something like a ‘forgotten password’ email, depending on the consumer. One of the pros of this approach is the ease of being able to see an indication of success within your platform at message or campaign level – in almost real time. However, it’s important to bear in mind that this is a post-click measurement. In other words, it’s tracking someone who has clicked through an email within the set time period, which could, of course, overlap with other channels. For example, if a customer finds your website via organic search, comes back via display retargeting, comes back again via social retargeting and signs up for your email before converting via paid search, this conversion could show up across all of your respective marketing channels.
Digging Into the DataMost of the clients I work with use Google Analytics, which is a great web analytics tool but not necessarily the best choice for marketing attribution. It has numerous flaws. The free version is subject to data sampling, and it doesn’t natively track cross-device. However, it does work better using Google login data, where someone is logged into Gmail across a number of devices. Nonetheless, Google Analytics can still give you some great data when you:
- Create email-based segments. You can then compare this to other channels and analyse onsite behavioural, demographic and product data.
- Look at Assisted Conversions. Where does email appear? This will help give you an idea of first touch and any touch contributions.
- Use the Model Comparison Tool. Similar to Assisted Conversions, your data can be broken down into models for analysis.
Uncovering Real ValueTrying to understand the true value of email can be tricky to say the least. Here are three effective options.
- Utilise native marketing attribution tools. Tools like Tableau and CUBED can help you efficiently tag your marketing channels, break down the data and look at a variety of different attribution models.
- Consider data matchback. Old school to say the least, this process involves exporting behavioural email data, such as email address, opens and clicks, and running this against your order data. It can give you an accurate picture of how many people on your email marketing list(s) actually converted – and how they responded to your last email.
- Try holdout testing. Not for the faint of heart, this strategy involves excluding a control group of subscribers from your campaigns so you can then compare their value to those that were included in email sends. This tactic aims to answering the following:
- How much revenue lift is being generated by sending this campaign?
- Would these customers purchase anyway even if we didn’t market to them?