30th March 2021
10 minute read
Fiction: A website can’t be optimised more than 1 or 2% unless it’s been really badly built
Fact: It can, we have, and this is how
We worked on a website which was already converting visitors to customers extremely well. We were able to increase conversion by a further 35%. We didn’t ask any customer questions and it took less than two days design and coding time. Lifetime value for each of these converted customers was at least £30-40k. I’ll tell you how we did it, but first…
I’ve had discussions with UX experts and optimisation specialists who told me that if a site can get optimisation gains of more than 2-3%, then the site hasn’t been built properly in the first place.
Most e-commerce sites – for challenger brands, where 1 or 2 products are being sold – go something like this:
They have a homepage, about page, ‘history/culture’ page – then the main customer sales journey; product page, add to basket, checkout. The argument goes that if you can optimise this process by a lot, then the people who first built the website have done a terrible job.
To some extent this is true.
The UX and UI of the ‘add to basket – checkout’ steps are problems which have already been solved. Take a look at how any of the big players do it – then do what they do. If you’ve got a Shopify website – it’s been done for you. Any improvements you make then will be minimal, and so are only really of benefit if you either have very high ticket items, or you have a large number of customers.
So if these steps are broken, or not in line with best practice, then sure, your previous agency may have made a mistake.
May. Your site may have been built five years ago, and best practice has since changed. Customers are a moving target – Apple Pay and Google Pay didn’t exist five years ago. People’s experience of e-commerce sites, and therefore their intuitive expectations, change as websites change. So it may well be that your previous agency was really good, but they just couldn’t predict the future. Your fault for not employing a telepathic agency.
The other thing to consider – ok, maybe you can’t optimise one part of that process by more than 1%. But if there are ten steps the customer has to make between landing on your site and paying for your product, maybe we can increase the conversion of each of those ten steps by 1.5%.
That will give you an overall increase of 16%. When you consider the increase in profit a 16% increase in turnover would bring to your business, that could be party time.
OK great. So if the site is a few years old, and it hasn’t been optimised recently, you can get some gains of around 10-20%. That’s probably the case for most businesses, to be fair. How many have already added Apple Pay to their site? Few. Immediate gains to be made.
But how do we get from 10-20%, to 33-50%? My contention is that all of the above is about creating a frictionless user experience, about optimising for the head – thinking, “where is the next actionable step” and making it as easy and as quick as possible to take.
To get further, bigger gains, you need to optimise for the heart. Getting more people to begin the sales process in the first place. We’re not talking about increasing the number of people at the end of the funnel, but the number we get in at the top. And importantly, how they feel when they start.
How your customers feel about your product is incredibly important. Compare these two situations:
Which one of these two purchases are you most likely to give up on mid-purchase if the website design is bad?
Think of a product that you really, really want. Often this is in niche, hobby products. For example, I really love card magic. There are countless inventors of card magic, people who spend their days and nights thinking about ever more obscure methods and principles around which card tricks can be built.
One of Britain’s greatest card magicians publishes his work very erratically. To buy it you need to first email him to check that he’s going to sell it to you. Then you paypal him – not from his site, there is no ‘buy now’ button – and wait until it’s ready, when he’ll post it out. There’s no notification, no confirmation, no ‘one-click’ purchase, nothing.
One of the world’s most prolific card magicians is only contactable via an SAE sent to a PO Box in the US. Hard to imagine a less direct way of purchasing from him.
I’ll happily jump through these hoops in order to get the thing I really want. That’s not to say that making it easier to buy wouldn’t increase their customers, of course. It’s just an extreme example to show that capturing the heart makes the head go through exercises it would never normally consider. It means we forgive a lot that we wouldn’t normally. Convincing the heart at the start of the sales process won’t just get more people into the funnel, it will, crucially, motivate more people to go through the funnel at every point. It will increase conversion at each step along the way.
This can have a huge impact on the overall conversion rate of the page. This is a point which is missed by running pointless A/B tests on different coloured buttons, placed in different positions. If I really want the product, I’ll buy it, regardless of whether the button is hidden at the top of the page and coloured in plaid.
Capturing the heart is difficult though. We often don’t know the reasons why we buy or want something. We have reasons we tell ourselves, sure, but they’re often not the actual reasons. The head thinks it’s the Oval Office, whereas it’s really the Press office, to paraphrase Jonathan Haidt. We make a decision without knowing why, and then justify it to ourselves afterwards. So asking customers why they’ve made a purchase and then optimising after that isn’t helpful, because they often don’t know why they’ve made the decision, or at least, not the real reason.
Tests, eye tracking and customer surveys are all useful approaches to finding the answers. Be careful though – these tests are often set up with our own presuppositions in place to begin with. We have a hypothesis and test it for a customer, and prove we’re right. This has some problems – as is beautifully shown in this video:
Another common problem is beginning a test with a limited hypothesis in the first place. To give a dumb example – if I give people a choice between an apple and an orange, and 60% of them choose the orange, then my conclusion back to the client is that ‘you should sell oranges’. If I run an A/B test with two options, and option B wins, then we conclude that B is best. But this ignores the fact that, in deciding to test A & B in the first place we could be missing a heap of other, better options. Bananas, for example, to carry on an analogy way beyond where it should have stopped.
One way around this problem is to run tests without anyone knowing they’re tests, and to make the tests as open as possible to avoid confirmation bias. Which brings me back to the start, and how we got increases of 35% on the conversion rate or the website with one small addition.
One of our clients is a large, multi-million pound rental organisation – they own the buildings, they do them up, they rent them out. We built their website to allow customers to book a viewing online directly with them, linking directly with their Google calendar, all self-contained and run internally by them. They have one and two bed furnished apartments in one of their buildings, and as part of the rent tenants get a bunch of different services and products included.
If I ask the client which of these products and services they think is important to their customers, they will probably be wrong, as they’re the client, not the customer.
If I ask the customers, they’ll say what they think they should say, or what initially springs to mind as ‘important in this situation’.
But if instead I don’t make any assumptions, but display on the page a table containing everything the customer gets as part of their monthly rent, we can then use software like HotJar to see where their mouse spends most of its time. It’s low cost, high returns, quick set up. We don’t need to set up users in eye-tracking software or set up a lab for a result from a limited set of people.
We can then see what they’re really interested in – where the mouse spends the majority of its time – and in this case, for a particular type of apartment, it was the type of furniture and quality of appliances which came with the apartment.
We then pull those into an image gallery on the specific apartment page, showing the right level of detail, with the right copy and the right photographs to really sell the aspirational nature of the products.
Result? Bookings for viewings went up 35%.
When we ask people what they’re looking for in a two bed apartment, nobody says ‘the washing machine’. They say the cost, or the location, or whether there’s a gym on site, or if they take pets. (I know; we asked).
But the type of furniture you have in an apartment, those little aspirational details, is hugely important when it comes to how we see ourselves. Which is hugely important when it comes to choosing between the myriad apartments there are out there.
It also becomes a comparison set for us to judge the apartment. We may choose a new apartment once every few years, so it’s hard for us to really know what constitutes a good or bad choice in the market.
So we use a comparison set of something we do know about to act as a proxy to help us make the decision. We know that BoConcept furniture is really nice; we know Smeg products are good ones. Knowing that information then casts the apartment in a good light, without it being an explicit statement. Presentation is as important as product – a meal will taste better on a china plate than on a paper plate (though again, context is key).
So how can you use this when looking to optimise your own offering? First, forget everything you know. Assume you know nothing, and then go about finding answers with an entirely open mind, ready to see your product range from your customers’ point of view.
Next, assume that what your customer says to you isn’t correct. It’s incredibly important to make decisions not on what people say they do, but what they observably do.
Create opportunities for your customers to show, not tell, you what they find important about your product, and then optimise from there. You may find big wins hidden in the unlikeliest of places, and, importantly, you won’t simply be testing the same things as everyone else.
Rob Dobson has been working in digital and building websites for 20 years. From designing and developing the world’s first internet bank in 1999 (smile.co.uk), he founded Northern Comfort in 2010.
See how we can help you.