The ultimate guide for using behavioural analytics and A/B testing to optimise website conversions

Content may be king, but data sits behind the throne and has the king’s ear. 

You want to be informed by data before you make changes to your marketing strategies. This is never truer than in the case of your website, which is a rich source of behavioural analytics and, therefore, a valuable insight into your audience’s interests.

What are behavioural analytics?

Digital behavioural analytics refers to the area of data analytics concerned with how people act on a site, application or other platform. 

One of the best starting points for iterative conversion rate optimisation is to trigger experiments based on user behaviour.

Phase 1: Create your launchpad

You need sufficient traffic and conversions for results to be statistically significant. You cannot base the success or failure of a particular form placement, word choice or page layout on a handful of data points. That’s a lot of work for little pay-off, and it could lead you to conclusions that are plain wrong.

As much as everyone wants to test what works and make data-driven decisions, the first phase you must concentrate on is to create, create, create. Build up your stock of marketing collateral and make your website the ideal launchpad for experimentation, using the best data you have up to this point. That may be real user data, if you have enough. Or, best practices that you’ve researched yourself, the advice of experts, and learnings inspired by your understanding of your target audience.

Phase 2: Install tracking software

Check if your website hosting software already includes tools such as tracked links and calls to action. You can also implement heatmap software like Lucky Orange or Hotjar. These tools show where users are looking on a page, what they click on, and so on.

Be careful not to overload your pages with code-heavy add-ons though, as this will impact performance.

Don’t think installing the software is enough. You need to be prepared to actually use the information it reveals, not let it gather dust while it weighs your site performance down. Someone has to keep an eye on the results.

Phase 3: Observe before you act

Saying that, patience is important, as long as you’re paying attention to the data. Allow the data to flow. View how interactions change over certain timelines, throughout one day, or over thirty days. Patterns should begin to emerge. You’ll be surprised at how people use your site — we’d bet it’s not how you expected. Probably, they’ve trodden their own path and muddied up the grass. It’s up to you whether you block that path off, or, better yet, transform it into a walkway fit for royalty.

‘Correct observation followed by meticulous deduction and the precise visualisation of goals is vital to the success of any enterprise.’

— Sir Terry Pratchett, author (and keen observer of human behaviour)

Phase 4: Plan your experiments

Now, apply the scientific method as you plan your experiments:

• Observation
• Hypothesis and prediction
• Experimentation
• Analysis

As an example, you observe on the heatmap that site users often linger on the homepage header and then click the button that reads ‘Chat to our sales team’.

You form the hypothesis that site users will be more or less likely to click on this button if the copy changes, and predict some word choice changes could improve the click rate, such as by changing it to ‘Book a free demo’.

Due to the fact site visitors often click this button, the data is likely to be meaningful and the return on optimising this important conversion opportunity is going to be worth the effort. So, you run a test by changing the button copy and measuring the results.

Once you have the data to make a like-for-like comparison, such as via a time-bound experiment, you can analyse the results. Based on whether the click rate has improved or not, you might decide to stick to the new copy or revert to the old one, or even try another experiment. The opportunities to improve are boundless.

A/B Testing

One way to speed up your testing and optimisation strategy is by using simultaneous tests. That is, rather than sequentially testing one feature after another, you use A/B or multi-variant testing to compare two or more variants at the same time, then pick a winner. Often, behavioural analytics informs A/B testing.

The value of A/B tests

The fundamental principle of A/B testing is to change just one thing at a time, so you control all other variables. Then, you know for sure that the change is what made the difference.

The advantage of simultaneous testing is that you automatically discount factors, like seasonality, which can skew your results. Meaning, if you test one landing page format in November and another in December, you’re going to be impacted by the seasonal dip in traffic that usually occurs on B2B websites during the winter holidays. That may give you a false impression that the first variant works best. But, if you A/B test that landing page in November, then you’re giving each variant the same chance at success.

What should you A/B test?

There are a lot of things you can A/B test. You should tie your tests to your marketing goals. Be mindful: you can’t run too many tests at once as that may mess up your data, so be clear about your priorities. If you’re focusing on traffic, then SEO tests where you try to optimise a page for certain keywords or performance might be your priority. Or you may run conversion rate optimisation tests to increase the number of leads filling in forms, such as by A/B testing landing pages. Tests based on user behaviour, which we discussed in the last section, are another great place to start. You can also test things for your own purposes, to drive behaviour and interaction. When you have no data to back up a change but you intuitively think it will be an improvement, use an A/B test to get that data before committing.

Here are some ideas of things you can test:

• Sections or pages on your website, in terms of the copy, format or imagery
• Landing page copy, formats and so on
• Forms (copy, number of fields, types of fields, and so on)
• Calls-to-action, links and buttons (location, copy, number of these on a page)
• Pay-per-click adverts
• Emails (subject lines, preview copy, body text, imagery, calls-to-action — there are a lot of things you can test on an email)

The never-ending cycle of tests

Running A/B tests is not always as easy as it sounds. Too often, again, companies don’t have enough data to make such tests viable. Or, they test the wrong things. Or lack the insight to translate test outcomes into actions — sometimes, a test outcome is not as straightforward as it seems. Maybe one variant does better by the number of clicks, but the other results in more buyers.

Or, companies fail at A/B testing because they don’t track test results and analyse ongoing post-test data. Why? Because A/B testing is like doing the laundry.

A/B testing software

You will require the right technology to allow you to run A/B tests. That is, a means to both run the tests and to track the results.

For example, you need some way to gauge when your test has reached the point of statistical significance. This may not be as simple as ‘one number is higher than the other number.’ Software should account for other variations and intelligently discount bad data, like bots.

Plus, ideally, you want built-in parameter settings based on the test setup. That way, when you create your test the platform you’re using suggests a test duration based on the criteria, and can notify you when a test has reached a point of statistical significance. Such information makes A/B testing much more accessible for non-data scientists.

As a HubSpot partner, we do the majority of our A/B tests using the Pro Marketing Hub because it provides intuitive testing setup and tracking systems. Other recommended tools are available, such as Optimizely.

How to run an A/B test

1. Choose your test subject, such as a marketing email.
2. Establish your test criteria, such as the subject line of an email.
3. Segment your audience.
4. Decide the test parameters (Such as a timeline. For example, you may run an A/B test on an email for four hours, sending it to a portion of your contacts, before choosing a winner and sending the rest of your contact list the winning variant).
5. Create your test with sufficient variation that it is likely to have an impact (As in, for an email, don’t make your subject line variants subtle. Do something completely different for each one).
6. Determine the metric you are measuring and the criteria for success (When it comes to email, that might be open rates or click rates. If you’re testing subject lines, a variant with a better open rate would be the success metric).

Then, of course, run the test, analyse the results and take that learning forward.

Optimising for conversions

You understand your users’ behaviour and you’re consistently testing variables on your site and other platforms to find opportunities to improve their experience. That is noble and right and good. But it’s hardly altruistic. On your end, you want to see a higher quality and quantity of leads come through the door as a result of your iterative testing efforts. That is, conversion optimisation.

Here are two ways to increase the number of inbound leads to your business:

1. Get more traffic to your website, OR
2. Improve the conversion rate at all the key stages of qualification, from lead to customer.

Now. SEO will get you more traffic. Backlinks from referring domains, pay-per-click ads, optimising content for keywords — all these strategies we’ve discussed. However, this alone is not enough. If your traffic stays the same, there are still opportunities to increase the number of leads and the percentage of that traffic that converts and becomes a contact. This is conversion optimisation.

Determine your conversion rate goals

Certainly, it helps to have a benchmark here. A benchmark is just a rough guide. If you’re miles below that number, then you know conversion optimisation is something you need to be working hard. If you’re hitting the mark or batting above average, then maybe it’s time to focus on SEO instead. For your reference, Unbounce and WordStream both report a median conversion rate of about two to five percent across industries. Unbounce says Software-as-a-Service (SaaS) companies get around three percent, but Finance sees a CR of over six percent, so FinTech’s may see their CR fall somewhere in the middle.

At Articulate, we run conversion rate optimisation efforts on many of our landing pages. As such, we have a ten percent CR of page views to contacts, and a whopping 25 percent CR for page views to submissions. Another thing worth pointing out is that the top 25 percent of landing pages get almost double the conversion rate of the rest, on average. These are the landing pages that have been thoroughly tested and optimised for conversions (and the content could also be more appealing).

As you can see, conversion optimisation pays off. So, observe where you are now, and set some specific conversion goals for each of your conversion opportunities. Then, test, test, test.

Here are three more tips to increase your conversion rate:

1. One form to rule them all

A/B tests are a great way to optimise things like forms for conversions, but they’re not the only strategy. Try making one master landing page form and using it on all your landing pages. Don’t use it for other forms, like ‘contact us’ or ‘book a demo’. These should be kept separate.

With the master landing page form, use queued fields so that when a contact reconverts on your site, you’re asking them for different information, like their job title or company size. Not only does this clue you into who your contact is, but it also reduces the number of visible fields on a form. Fewer fields mean a lower barrier to entry, so more people are likely to fill in the form in the first place and to convert more than once, grabbing multiple downloads in a Netflix-style binge session.

Once someone has filled in a form on your website, they’re much more likely to do it again, especially if you signpost the next chapter on their journey with clear calls-to-action.

2. Make your blog a conversion powerhouse

If your inbound marketing machine is running well, your blog is the primary source of traffic to your site, so it stands to reason that you should focus your conversion efforts on it.

Add CTAs to blogs and test their placement, subject matter, copy and so on. Put conversion opportunities in the sidebar. Add a ‘recommended reading’ section to every article. Include the option to download articles as a PDF, or as an audio file. Give people the chance to subscribe to your blog. You could do the same thing for webinars. Leave no stone unturned. Pack your blog with goodies.

Top tip:

Click the CTA at the end of this blog to see how effective this can be!

3. Optimise for the mobile experience

Finally, many B2B businesses operate on a desktop-first basis when it comes to their website. Most visitors (at the time of writing) do probably come to your website on desktop, but — mobile is ever on the rise.

We’ve all had that experience of a giant pop-up covering the screen on mobile, with the ‘close’ button cut off at the top so you can’t even get rid of it. What do you do? Leave the website. Disaster.

So, put together strategies to target conversions from mobile users, like mobile-specific CTAs or simplified pages with conversion opportunities higher on the page, so you don’t have to scroll too much.

One decision after another

The job of anyone in a powerful position is not to make perfect decisions but to make decisions. Full stop. Indecision is the enemy of iterative optimisation. Decide, then track, observe, test and, if needed, make new decisions. No experiment is a failure, merely an opportunity to learn.