Bryan Eisenberg certainly thinks that ‘Conversion Optimisation’ is the best thing since sliced bread. In fact he’s just left his company to focus on being a marketing speaker covering this ‘critical’ aspect of marketing and recently demonstrated his thinking through his excellent Search Engine Strategies keynote “21 Secrets of Top Converting Websites”. But what is ‘conversion optimisation’ and how does this fit in internationally? I’ve gone into some depth on Bryan’s views in this article – because I found them particularly useful myself including making some changes to www.webcertain.com. But I would like to point out that the meaty comments here are all Bryan’s – and the smart ass less useful ones are mine! My number one tip is ‘Buy Bryan’s books!”
Bryan is certainly a data-lover. Unlike those who in a recent survey said that they would rather manage without their partner for a week than their laptop – Bryan would certainly sacrifice his Mac but don’t ever try and take the data away. Not only is data the basis of conversion optimisation it is also the key to getting senior management support – always a challenge for those working in the online marketing space. He revels in being provocative with his audiences at conferences with statements like “You don’t have a traffic problem” at an SEO gig where the principal focus is around traffic – with yes a casual nod towards conversion.
Get aligning those objectives
Bryan points to the relatively low conversion rates on most sites which typically run at 3% or less and yet marketing teams focus on expanding traffic rather than digging into site performance so as to convert more visitors into customers. To justify an investment in starting to test, he suggests calculating the returns that would be made in the bottom line by making a small improvement in conversion rate and to sing Bryan’s own mantra – and the title of one his books – “Always be testing!”
But becoming data driven is not just about dashboards but more along the lines of Demming’s ‘total quality management’ as improving quality will actually reduce expenses. Says Bryan, “All the big brands you can name from Google and Dell to Xerox and IBM have always focused on total quality management – that’s why they’re as big and successful as they are.” Amazon, for instance, didn’t get to become a global brand by selling books, they did it by aligning customers and business objectives for years and by paying attention to detail and trimming the fat. Amazon not only tests things that work, but things they’ve tested many times before just to see if there or further improvements to gain or if their customers behaviour has changed. Dell tested the words ‘learn more’ against ‘help me choose’ with the result that ‘help me choose’ drove more conversions through the site. They’re all doing it because it makes a difference.
Getting a champion is a good idea!
But where do you start especially if you’re not a global player and have more limited resources. With passion, Bryan stresses, “Just test what you can test – start somewhere, anywhere. If you can only test headlines then test headlines.” It’s easier to double your conversion rate than to generate more traffic. The first step is to check how much of your budget is actually invested in testing and whether that really makes sense. It used to be expensive to undertake effective testing on a site but that’s no longer the case. With product’s such as the Google Website Optimiser coming on to the market and offering facilities absolutely for free – there are no longer any good excuses not to get on with it.
Key is not thinking that you have to make a list of all the tests and then run them thinking that you’ve tackled the issue. The secret is not the first or the second test, but building testing into the company processes and making it a way of life. But you do need a champion who is going to live, breathe and own the process or it will be difficult to get things off the ground. Look for the right creative resources, which might possibly be those you have internally, but it’s more probably wise to go and get external designers to work on your testing to get some degree of objectivity.
Who are you trying to persuade?
Perfection is sometimes a barrier to getting people into testing. (Perfection is sometimes also a barrier to SEO or indeed anything!) People often start by trying to make their planned test absolutely perfect rather than just getting out and getting a test in place. Equally “Don’t do slice and dice optimisation,” says Bryan. What he means by this is that you shouldn’t look at all of the individual components of a page taking each element and testing it. You need to look at the whole page and test its components within the context and role they play within the page and ultimately the page design needs to ‘hang together’ dude as a consistent whole. If getting testing started on the main site is really tough, then you can begin by picking some pay per click terms to get things going. But don’t pick the biggest and most important or you might be taking too big a risk with your first experiment.
The secret sauce of the Eisenbergs – Bryan works with his brother Jeffrey – is called ‘Persuasion Architecture’ which Bryan and Jeffrey launched, developed and trademarked some time ago. Put in the simplest of terms, the persuasion architecture approach consists of answering three key questions:
1. Who are we trying to persuade?
2. What action do we want them to take?
3. What action do they want to take?
What type are you?
Notice anything? There is a noticeable conflict set up in the questions themselves which if you’re sharp you’ve already spotted? Does the customer actually want to carry out the same action you want them to! Now I’ve not seen the detail behind the persuasion architecture idea – but I’m guessing that this conflict between what the customer wants – and the marketer – is the basis for the Eisenberg solution. Part of the solution is also that people are different. No they really are – I’ve noticed that.
Bryan refers to personality typing systems – such as Myers-Briggs – to come up with four core personality types which it is wise to accommodate in your marketing planning. (I have to say, as a user of the Myers-Briggs approach for some years, it’s a relief that there are only four types – rather than the 16 in the MB system and for those who’re interested I’m an ENTP). These four personality types are also connected to Bryan’s answer to my questions relating to how the conversion optimisation approach can be applied internationally which I’ll go into in moment.
Working with segments or personnae
The four basic personality types which Bryan describes as having a significant meaning for web marketers are:-
The names of the different types are pretty much self-explanatory – but in his presentation Bryan proves, with the use of eye-tracking studies from the famous usability guru – Jakob Nielsen – that these different personality types do actually demonstrate distinct behaviours provoding proof of concept. Methodical types needs to become experts themselves, the spontaneous love pretty pictures, humanistics adore ‘about us’ pages and reviews whereas competitives make quick decisions and jump to conclusions.
Knowing these different types and behaviours is significant for people working in online marketing because knowing how people gather information informs how they should be addressed by websites. Typically online these groups are known as personae – though marketers for years have been calling them segments. (Same thing – different name. We can’t have web activities sounding like they were invented before the web can we?)
Thinking about conversion internationally
These personae are manna from heaven for multivariate testers – that’s the people who make many many changes to test a site very quickly – sometimes testing millions of elements at once. Knowing the types means they can see who they match up with different elements of a page – and no, the fact that there are four key types does not mean that you have to re-launch four different versions of your site – thankfully. What it does mean is looking at the site four times to see what messages and impacts will be delivered to the different types of visitor.
As we are just in the midst of making some changes to the WebCertain.com navigation system, I threw this idea at our marketing team and, interestingly, it did make a difference to what content we showed at the top and how we described things. Thanks Bryan. (These changes launch in a couple of weeks).
Offline, I asked Bryan how he saw these personnae relating to conversion optimisation à l’internationale. Bryan’s view is that the same personnae apply worldwide but they may well differ between markets and industries so you have to do your research. In fact, he points out that the Myers-Briggs types already accommodate these differences in that it is already known that the types vary by nationality and culture. What this means is that a water pump manufacturer may actually find that the personnae are pretty similar worldwide (more typical of business-to-business than business-to-consumer). They are less likely to have ‘spontaneous’ web visitors than ‘methodicals’. A fashion retailer targeting consumers will likely spot many more differences.
Persuasion is not an event
Bryan’s presentations are full of great examples. In one stunning case, he shows a graphic which was tested resulting in a very small change which – he says – generated over $25 million of extra business for client. (I couldn’t help wondering at this point how and how much Bryan had actually charged his client but didn’t dare ask. Percentage share?) Both macro and mico actions are important. You not only need to understand the micro actions – such as why aren’t they clicking my jolly button – but also to look at the macro elements which is more to do with connecting the marketing messages with site journeys and processes. So, for instance, if insurance company Geico, runs a compaign targeted at ‘competitives’ where they are able to fill in the minimum of details quickly, don’t follow that up with a contact form with lots of additional mandatory fields.
Above all persuasion is a process not an event. Always ask yourself on every page why your customers should want to buy from you. One of Bryan’s books is called “Waiting For Your Cat to Bark” – so I was not particularly surprised to hear web-user-customers described as bloodhounds. There are two reasons why Bryan uses this analogy. One is that web users are exceptionally goal oriented – this is something which has been shown by study after study which I’ve personally consumed. But on top of that, usability folks like to talk about ‘information scent’ a lot – which fits the bloodhound well. Information scent was first understood by the Xerox team at the PARC or the Palo Alto Research Center.
“Watch out” page load speed is coming
55% of all people leave by the second click and 17% by the third click. With conversion rates typically running at no more than two to three per cent it stands to reason that action is needed. Says Bryan pointing at the audience in a rather ‘The Apprentice you’re fired’ manner, “If a keyword is relevant, the keyword doesn’t fail to convert you do.” So now you know.
Usability testing doesn’t necessarily mean long set-up cycles and expense either. By getting five testers from an organisation such as www.usertesting.com to run through your site trying to meet certain set – you can gain more information than you can from all of your web analytics data for a month. Now that’s some claim! These tools are also very powerful and easy to use.
As we all know, Google has laid out its stall for 2010 with very clear messages that page load times will be more important for both SEO and pay per click. They are not saying exactly what this will mean in practice but the “watch out” message is coming across loud and clear. Naturally Bryan is all in favour of this approach and users really DO NOT like slow loading pages wherever they are. He recommends looking at images on your site as this is a relatively easy fix and yet can consume lots of download time.
Landing pages have an anatomy?
You can do a speed test on your site at www.websiteoptimization.com – the speed tool is here. You need to bear in mind that interaction, a desirable for web marketers, means that users need to see responses in under one second.
Bryan recommends using smush.it – which is now a tool within the Yahoo developer network – to reduce the images. It quickly provides a reduced size image with a link to the file already appropriately ‘smushed’ ie reduced to an acceptable web size.
Go figure – landing pages have their own anatomy. Is it serious doctor? Bryan explains that you can pretty much predict the common elements of pages and identify the main sections. There’s usually a logo, a headline, an offer, some copy and calls to action. (Actually the calls to action are usually missing or below the fold in my experience).
Page components and their features
But once you know the anatomy you can start to structure the way you’re going to test including looking at whether the page components play their role well – or are they just freeloaders? Ask yourself these questions:
- Is it relevant?
- Is the quality good?
- Where is it on the page?
- What’s it next to?
- What stands out?
The features of the page components you are testing should look at:
Think about building confidence
Buttons are the easiest to test – and can make significant differences. And as for forms – just keep on testing them – that’s what all the big boys do and with good reason. Customer needs change, the competition such as forms and other competing websites improve and we should never forget that users have a choice. One of the things that influences their choice is the website’s functionality and it’s ease of use.
When testing forms the things to consider are the size’s of parts of the form including buttons and fields, the field’s and what they do, the field labels, benefits shown on the form page (yes there should be some) and confidence building materials on the page such as testimonials or client comments!
Every detail in a page matters and it is particularly important to ‘build confidence’ in the page using third party contributions, user-generated content, testimonials and resources to answer specific questions. Oh and don’t forget to look professional! It’s best if every page wears a suit and tie! Most especially give assurances directly at the point of action.
Looking at things from different angles
When Bryan says ‘speak the right language’ he really means use the language with which your customer has empathy in terms of tone, formality, technical level and so on. But he wouldn’t disagree with me at all when I say that when working internationally it’s really not a good idea to skimp on the content quality or expect the user to put up with reading only partial content in their own language and then having to go to a English language form. Whoops.
Pictures of the product matter. Wow does Bryan really have to say that surely that’s obvious – er no it isn’t in fact. The quality of that picture counts but does it answer the potential buyer’s questions and anyway what are those questions is the general picture pitfall. it might be, for instance, different angles of the product if, say, it was a boot. Or you might get better sales if you have pictures of all the different colours you sell that attractive polo shirt in.
“Test, test, test”
Sorry Flash fans, Bryan isn’t a member of your club. He reckons that changing that animated flash banner to a static one with bullets can actually reduce the abandonment rate. You should also avoid using we-we too much in the content of your website – the Eisenbergs have once built a tool which pulses electricity through your chair if you’ve overdone it. (Sorry Bryan – I have a problem with this one because I use ‘we’ all the time in my copywriting but as the voice of the client as in “We need to get our act together on international SEO. That’s a challenge for many but now you’re in the right place.” Never forget to show off the character of your brand – though you might find it easier to work with third parties to give you clear objectivity on this one.
The key point of everything Bryan says is “test, test, test” – it’s the only way to get close to your customers’ needs – and that includes those in other countries. You have to adopt the testing mentality if you’re ambitious and want to be the next Amazon. I’d accept the next Amazonette – but still.