Marketers have long had a tendency to invent new terms and then expect everybody else to know what they mean. Over the years we’ve moved from SEO through to Inbound Marketing via Multi-Channel Attribution and Real-Time Bidding (to name but a few).
We’re all a bit guilty of this, so I thought I’d try and explain some of the more common CRO terms without too much jargon. As this is such a huge subject, I’ve also taken the opportunity to link to various other resources on the web that have already done a great job of explaining some of these concepts.
Conversion Rate Optimization (or CRO)
Most of the time, your website will have a purpose. It may be to sell stuff, or to convince somebody to give you their email address so you can sell stuff to them at a later date. It might be to get somebody to watch a video, or even pick up the phone to give you a call. If a visitor is kind enough to do any of those things, that counts as a conversion.
The conversion rate is simply the percentage of your visitors who go on to complete the desired action. If you have 1,000 visitors to your site, or landing page, and 10 people go on to buy something/part with personal information, then you have a conversion rate of 1%. (I kept the maths simple here for my own sake.)
So, CRO is the process of improving, (optimizing) your website to increase the percentage of visitors that convert.
One of the reasons CRO is so important, is that not only do your users think you’re lovely, as you have a wonderful, easy-to-use site that helps them achieve their goal, but it’s much more cost-effective to convert your existing customers than it is to attract new ones – increasing the return on your current marketing costs.
There are many tactics and techniques you can use to increase conversion rates, some of them are explained below, but if you want a definitive guide to CRO than you could do worse than read this guide by QuickSprout
The staple in any Conversion Rate Optimisation toolkit, often called a “split” test. This is where you test version “A” of a web page vs version “B”. This is usually done through sending 50% of your traffic to one version, and 50% to the other.
Sometimes this will just be a simple test, testing one hypothesis. Say you’ve been looking at your web analytics and you’ve realised that visitors to one of your landing pages don’t convert as well as visitors to the rest of your site. You look at that page and see that users have to scroll down before they can see the “Buy” button. You hypothesise that by moving the button up to the top of the page, more people will see it, and therefore click on it.
So, version A is the existing page, Version B (the challenger) has the “Buy” button higher up. Run the test until you have enough traffic volume to make it statistically relevant, and the version with the most conversions wins!
Version B then becomes the control against which to run future A/B tests.
It’s really important to make sure you have a solid strategy behind your A/B testing. Otherwise you could constantly be running small tests that don’t have a big enough impact on your revenue.
Here’s a great resource from ConversionXL about making the most of your A/B Tests
Basket Recovery is closely linked to Shopping Cart Abandonment, and is the act of trying to recover the sale once the user has left your purchase process.
There are various methods used for basket recovery:
- Email – an automated email can get sent as soon as somebody leaves the process (or at some pre-determined time). These can often be incentivised – i.e 10% off if you come back and buy – some good examples here.
- If your product is slightly more complex and maybe of higher value, it may be worth routing that information into your outbound call centre. A follow up call could often rescue the sale, or at least gain insight into why somebody left
- Overlays – although these can be seen as a bit of a nightmare, and if used in the wrong place can often get in the way of a sale – many brands have seen success with an overlay once a user leaves the site a specific point – again, these could offer a time-sensitive incentive
- Re-marketing. There are various ways you can re-market to people who have been to your site, and left at a certain point. Google’s re-marketing functionality is particularly straightforward – You have to give a few details to get this guide from Wordstream, but it’s a handy guide.
Behavioral targeting uses technology which allows you to serve different adverts/content to various customer segments.
You can serve ads to online visitors across multiple websites, based on the user’s previous actions such as websites visited, searches made and other ads clicked on.
This allows consumers to receive ads that are more relevant to their interests and behaviour, and increases the likelihood of a conversion.
Bounce rate is an oft misunderstood metric. Simply put, it’s the measurement of people who go to a specific page on your website, and then don’t visit another one. It doesn’t take into account the time a user spent on the page, or whether they interacted with it in any way. In this way, it’s a metric you need to be careful of.
For instance, if you have a campaign where your main aim is to get prospective customers to read a blog on a particular subject, that user could go to your site, read the entire blog, copy the link and leave the site to go and share it with their friends, but they will count as a bounce, even though they did exactly what you were hoping for.
Our own homepage has a seemingly high bounce rate – but a lot of our clients come back to our homepage to log-in to their account – so they hit the home page and then go straight off to the SessionCam Dashboard, which is on a different sub-domain. Therefore, they too count as a bounce.
Having said this, a high bounce rate for a page which gets a lot of your traffic, could also be a sign that all is not well with the page. It’s also commonly understood that Google take engagement metrics such as Bounce Rate into consideration when ranking a website.
So, if you feel you have an issue and want to reduce your bounce rate – there’s some good tips from Search Engine Watch
Call to Action (CTA)
A Call to Action, or CTA is simply anything that enables a user to DO something. These are usually buttons, links, copy or images that encourage the user to:
- Buy Now
- Play a video
- Download a PDF
- Complete a contact form
- Sign up for a newsletter
- Pick up the telephone!
- Anything else that counts as a way for a customer to convert!
CTAs are significant in the Conversion Rate Optimization process, as this is arguably where you want your user to focus – the main purpose of your page is to allow your user to act. If they don’t see the CTA, or there are too many on a page, this can cause your potential customers to run away.
Common tests include:
- CTA colour
- CTA Copy
- Size of CTA
- Position of CTA on page
- Number of CTAs on page
- Link vs Button
And so many more. Wordstream suggest a few best CTA practices in their blog.
If you have an e-commerce site, you’ll likely have a shopping cart or basket. If a visitor puts something in their basket/cart and then leaves before parting with their money, then they have “abandoned” the shopping cart.
This is rarely done in real life (although I have been known to abandon a half-full shopping cart in Asda once. It was near Christmas and the pressure got to me). However, it is all too common in the online world.
A recent report stated
“Approximately $4 trillion worth of merchandise will be abandoned in online shopping carts this year and about 63% of that is potentially recoverable by savvy online retailers”
However, it’s worth noting that not everybody who “adds to basket” ever intended on buying the product in the first place This report also shows that a significant amount of shoppers who have abandoned shopping baskets were planning on actually going in-store to complete.
Retailers can reduce the rate of abandonment and increase conversions through identifying the reasons customers are leaving, making the purchase process as simple as possible and providing the user with all of the support they need to go on to purchase.
Also – all is not lost even if somebody does leave their shopping cart abandoned in the middle of the virtual aisle – there is always basket recovery.
Understanding your conversion funnels is essential to CRO. A conversion (or sales) funnel is the path by which your customers travel to get to the end goal – the conversion. So, at the top of the funnel you may have your PPC ads, followed by a landing page, a basket, a checkout form and a “Thank you for purchasing” page as the final step.
Conversion funnels can be as long or as short as you want – they may be a simple two-step process – a visit to a blog page followed by a newsletter sign-up.
However, by clearly identifying the pages/steps the user goes through on the way, you can start to identify where the funnel has “leaks”, and try to re-engage your potential customers.
After setting up conversion funnels in SessionCam, one of our clients identified they had a significant drop-off rate on their “Find a Store” page. By viewing the sessions of the visitors that had dropped off, they could see that the postcode field was causing a lot of issues for people , either they were entering their postcode in the wrong format or entering a town name that was not recognised. By changing the validation on this page, this could be fixed quite simply to stop users “leaking” t this point.
A good place to start with some basic funnels is in Google Analytics. Kissmetrics have a produced a really handy introduction to remarketing.
This is a metric that is often confused with Bounce Rate – but is definitely not the same. Whereas bounce rate measures anybody who just hit one page on your site – exit rate looks at the amount of people leaving from any individual page, regardless of how many pages they’ve visited beforehand.
Therefore, pages with a high exit rate can indicate an issue with your customer journey, especially if they are pages in the middle of the conversion funnel. You would probably expect a high exit rate form the “Thank you” page, but would be quite concerned if your basket page had a 100% exit rate.
Website heatmaps give a lovely visual display of the actions of a selection of visitors to your site. You can see where people moved their mouse, what links they clicked on and how far down the page they scrolled.
By using this aggregated data, as opposed to just looking at an individual user journey, you can quickly highlight actions multiple users are taking, especially ones you might not want them to!
We saw one client who looked at their click heatmap on the checkout page, and found that the “Click and Collect” option was being repeatedly clicked on, even though this wasn’t actually a link.
This could easily be remedied, by either making that area clickable, or by changing the formatting so that users didn’t assume that it was.
We recommend using filtered data to get the most out of your heatmaps. This can then show you how different customer segments interact with different pages.
We’ve written another blog that shows you how you could get the most out of heatmaps.
Your hypothesis should form the basis of every test you conduct. Once you’ve gathered your data through VoC research, Analytics data, Funnel analysis etc, you should then have formed a valid hypothesis on which to test. So, rather than saying “I want to test the Call to Action Copy”, your previous research will have indicated that customers aren’t interacting with your call to action button. Your hypothesis should therefore be “Customers aren’t interacting with my CTA. I hypothesise that if I move the CTA button higher on the page, more people will click on it”
Ideally, your hypothesis would also include a measurable element – so “10% more people will click on the CTA if I change the copy”
Once you have your hypothesis, you can then measure what success looks like.
More detail on how to form a valid hypothesis from MOZ
Although this may seem obvious, it’s still worth spending a bit of time on the importance of landing pages. Many people spend a large amount of time optimising their homepage, as this is seen to be important within the organisation. I’ve actually worked with one company who were constantly running homepage tests, but after looking into their analytics, only 7% of their visitors ever even saw the homepage.
Every page on your site that has a desired outcome is a landing page. As are all of your campaign landing pages. (Don’t have separate landing pages for all of your campaigns? You probably should.) The more you can provide the ideal landing page for all of your customer segments, the better your conversion rate.
I would go into more detail about how to optimise your individual landing pages, but Unbounce have already done such a great job, you should read their 101 tips on landing page optimization.
This is where your A/B test gets a little more complicated. As opposed to just testing one element on the page, or one page design against another, MVT allows you to test Multiple Variations on a page. So, rather than just testing one element against another, i.e two versions of your Call-to-Action copy, you can see how that change works in conjunction with other changes, such as an alternative form length, or new headline copy, and identify the best-performing page overall.
The benefits of this is that it allows you to test multiple elements in one go, avoiding the need for several A/B tests. Also, a traditional A/B test might let you know that changing a button colour to green improves conversions against the original blue button, however you may never find out that the original blue button may have performed even better if used alongside a different headline.
The downside to MVT is that you need considerably more traffic to get a statistically relevant result. Rather than splitting your traffic 50/50 as you would in a simple A/B test, your traffic is probably split into at least 8 variations, so bear that in mind when planning your test.
Personalisation is where a brand strives to deliver an individual experience to each of their customers.
As Econsultancy’s research shows, 74% of marketers know that personalisation increases customer engagement. However, the majority of brands still aren’t using this to the extent they would like.
Although providing a unique experience for every one of your customers may seem out of reach, there are still things you could consider, a lot of which can be automated (This may not seem particularly “personal” but still delivers a more relevant experience for the user)
- Email personalisation – not just using their name in the Subject Line, but actually delivering content depending on what you know about them. This may be their purchase history, or articles that they’ve clicked on in the past
- Web personalisation – serving different content to each customer segment
- Behavioral targeting (see post)
- Website automation such as “Recommended products”
A lot of this depends on making sure your data is accurate and you really understand your segmentation. Sometimes bad personalisation can be worse than no personalisation at all!
Qualitative and Quantitative data
Although both qualitative and quantitative data are terms that have been around for quite a while, this is what we mean in relation to user research:
Qualitative data is gathered by interactions with real people, often taking into account personal opinions, and tends to be subjective. Qualitative research can take the form of:
- In-depth interviews
- Group discussions (focus groups);
- Diary and journal exercises
- oice of the Customer (VoC) feedback
Quantitative data is number driven, looking for the actual data or statistics to support or dismiss the hypothesis. Examples include:
- Survey data
- Web Analytics
- Net promoter score (NPS)
It is necessary to use both kinds of data when looking at CRO. You could use them to support each other – i.e.
- A customer complains via live chat that they find a form field impossible to complete
- By looking at conversion funnel data, you can identify the scale of the issue and see if that could have a significant impact.
Or, the other way round:
- Your analytics data shows that you have a high exit rate on a particular page
- You use session replay to analyse visits to that page and see why and how they are leaving. This then gives you the insight for your testing hypothesis
Customer segmentation is the practice of grouping customers based on similar characteristics, allowing you to better understand their needs and therefore deliver a more personalised experience.
A traditional way of segmenting would be based on standard demographic data – looking at gender, age, income etc.
This has moved on considerably in the last few years and although demographics still have some part to play, marketers have more refined ways of looking at customer behaviour.
Some examples of segments you may want to consider are:
- Traffic source (Customers who respond to an ad in Google have had very little information, so they may need more guidance once they land on your site)
- Previous interactions (Users who have already downloaded a White Paper from your site may have a greater propensity to buy from you as they already trust you)
- Customer lifecycle – where are they in the customer lifecycle – customers who are looking at competitors will be looking to see what makes you stand out.
You can also segment users by their value to you, or by geography, what device they use – the list goes on. Find the segments that have the most distinctive group behaviour, and you should be able to tailor your content to them nicely.
Session Replay, or User Replay, is software that allows you to record actual visits to your website. By putting a simple tag on your website, you can record visits (sessions) and then play them back to see exactly how your user interacted with the page.
You get a video style recording which lets you see exactly what the user saw, where they moved their mouse, where they clicked and even what they entered into form fields.
You probably don’t have time to watch every visitor to your site, so the best way of using session replay is to filter the visits to the site in some way you define. You could filter your traffic by where they came from, or by whether they were on mobile or desktop. You could also identify pages with a high exit rate and view a selection of sessions to see why people are leaving.
Again, if you’ve set up funnels, you could identify the page with the highest drop-off and replay those sessions to get some insight.
Combined with other analytics and testing tools, session replay can provide some really powerful insights.
User Experience (UX)
Having spoken to a couple of experts, and scoured the web for a definitive answer, it seems that people still can’t quite agree on the exact definition of User Experience.
When people speak about usability, they often use the term user experience to mean the same thing. However, I would suggest that Usability is about making something easy to use. Usability focuses on allowing a web user to accomplish a goal as simply as possible.
This differs from this definition of User Experience:
Every aspect of the user’s interaction with a product, service, or company that make up the user’s perceptions of the whole. User experience design as a discipline is concerned with all the elements that together make up that interface, including layout, visual design, text, brand, sound, and interaction. UE works to coordinate these elements to allow for the best possible interaction by users.
This separates UX from usability, in that usability is about allowing people to achieve a task within a product or website. So the difference here is all about the “experience”, and takes into account the feelings or emotions involved.
There’s a nice article here which compares the “experience” of using a Blackberry to using an Iphone…
Voice of the customer (VoC)
One often over-looked way to see what users think of your site is to ask them! Customer perceptions can be key, and although we know that not all users may know exactly what they want, their comments can often provide insights for further research.
Common VoC tools you can use are:
- Live Chat – easy to implement and allows your customers to engage across your site – you can see where they are having questions and what issues they are facing in the moment
- Surveys – there are many simple survey platforms which you can use to gather customer feedback. Ask your customers what they think of you and why!
- NPS – Net Promoter Score, this is a measurement calculated based on responses to a single question: How likely is it that you would recommend our company/product/service to a friend or colleague? Many organisations claim that a company’s NPS correlates with revenue growth
Fundamentally, CRO is concerned with the customer experience – so communicating with the customer directly seems like a bit of a no-brainer…
Well hopefully we’ve explained a few of the common CRO terms, but we’ll keep adding to this blog, so if there’s anything you’d like us to cover – please just let us know!