October 07, 2010 The impact of email activation on conversion rates Tracy Godding, Head of UX Research

The impact of email activation on conversion rates

You may remember that a while back, I put together a screencast on lazy registration, demonstrating how we'd achieved it in the Madgex Job Board Platform. Building Lazy Registration into any kind of webapp is a lot of work - but it's a very good idea. After all, if you take a complex, multi-stepped process and turn it into a rapid, single step process, there's no doubt that you conversion rates will go up.

Even though this is obvious, there's very little published data on the kind of uplifts you can achieve by implementing lazy registration properly - i.e. taking out all of the unnecessary steps from a sign-up process. In this post we're going to set this right by sharing some data that one of our clients, Haymarket, has kindly let us publish.

This article focuses on email alert creation and the impact of mandatory email activation on the conversion rate. In other words, we're going to look at what happens when you send users an email and require them to click on a link before they can do anything else.

Email alerts aren't the sexiest part of any webapp, but they are extremely important. The conversions you get from email alerts can easily be double (yes, double!) what you get from organic search. In our case we're talking about job application rates - in your case it may be ecommerce transactions, bids, or indeed any other kind of conversion. The reason email alerts give great conversion rates is quite obvious when you think about it - a user has explicitly specified their search criteria and has asked to be notified only when new matching items appear in your inventory. It's not surprising that they are more likely to take action from an email alert than a random person surfing in from Google.

With this in mind, we decided to move email alerts outside of the registration barrier in our new V3 platform, so users can create (and delete) them in a single step, without needing to register or verify their email address. This delivered huge uplifts for Haymarket, who were upgrading from a bespoke platform that included an email activation step in the email alert creation journey, as shown below:

Before and After - Email Alert Creation Process

Lets look at some hard data demonstrating the value of this change on Brand Republic Jobs. (If it's not a site you're familiar with, check out the Brand Republic page on Alexa.com to get a sense of the site's size). I think you'll be able to guess the date of launch:

Brand Republic - Email alert creation rate - before and after

You'll notice that there are no labels on the Y axis of the graph - unfortunately this data is confidential and we're not able to share it with you. One of the things we should emphasise here is that this is not a scientific study. With before-and-after data, you always get numerous uncontrolled variables. There are lots of things that are different between the old and new platforms, and besides, in the natural world there's lots of stuff that varies in the passage of time. If we'd have wanted to go to town, we'd have run an AB test, removing all the uncontrolled variables and serving two different versions of the user journey to randomly selected users in the same window of time. However, we really didn't feel the need to do this - we knew the new platform would deliver an uplift, and our remit was to launch it quickly and efficiently.

So lets not get carried away here - this is indicative data, but it's not the result of a rigorous scientific study. That said, a T-Test revealed the uplift on Brand Republic Jobs to be highly statistically significant (P<0.001).

Since that's just one site, let's take a look at the effect of the exact same platform upgrade for Campaign Jobs Asia. It's a smaller Haymarket site, servicing a different audience in a different geographical location.

Campaign Jobs Asia - Email alert creation rate - before and after

Above you can see a very similar uplift. A T-Test also revealed this to be highly statistically significant (P<0.001).

There are of course certain types of site where the verification of email address ownership is critical to a site's security, so you should be careful when considering this change. However, if you are able to do it, the graphs above show you could stand to benefit greatly.

While this is a useful tip in itself, you should bear in mind that there's a far bigger picture issue here: you really shouldn't be gaining this kind of insight from blog posts like this. Instead, you should be running your own iterative research and design programmes for your own sites. If you're not gathering and actioning these sorts of insights on a regular basis, then you're always going to be playing catch up.

Of course, running this kind of programme is expensive, and if you can't meet the costs, outsourcing is one way to go. For example, all Madgex customers benefit from our iterative research and design programme which feeds into our regular product release cycle. In other words, if you buy one of our White Label job boards or CV / Resume Databases, you'll find that for the life of the contract, it'll get updated every few months with User Experience enhancements - at no extra cost.