After watching jealously as all the SEO types in the office were swanning off to BrightonSEO and The Content Marketing Show, SiteVisibility delivered a free conference for geeks like me in the shape of MeasureFest. The whole day was devoted to CRO, Analytics and the wonderful world of business intelligence. What was so great about it? In a nutshell, great insights, interesting speakers and awesome tools.
But you don’t want it in a nutshell do you? Let’s go through the talks one by one:
He started up by setting out the challenge that faced him at Distilled – to make the company awesome at CRO. He was pretty honest about the downsides of this task: CRO is risky and it’s hard to develop expertise without live projects to work on.
They may sound like obvious statements, but they are really fundamental to getting into this area. If you want to work on live sites, inevitably you will have projects that don’t work out as expected, and communicating that to any client is essential. Similarly, even reading everything you can about CRO, doesn’t make you an expert on CRO. Not that you shouldn’t offer CRO services – but when it takes 10,000 hours to be a true expert, that’s hard to master.
It may sound like this was Paddy’s way of making Distilled awesome at CRO by scaring everyone else out of the game, but he followed this up by sharing with us his framework that can be applied to any CRO project.
You can get a lot of the steps from looking through the slides, but here’s some extra detail to help you make sense of it all.
- It helps you to understand what you really need to change on the site to (theoretically) affect change. Without understanding what is going wrong currently, how can you make sensible decisions about changes?
- It helps you to show value. If you want your client/bosses/future clients to think that you are the Most Amazing CRO That Ever Lived (MACROTEL) and they really should give you lots of money as a result, you’ve got to be able to prove it with cold, hard facts. Ignore this stage at your peril.
The other key takeaway from this is that you’ve got to go beyond easy data, talk to sales staff and customer support and don’t rely on what you can see in Analytics.
If you did stage 1 right, you’ll have a clear idea of what you need to do to make your site awesome. To cement your MACROTEL status, you’ll need to need to be as scientific as possible.
But it isn’t all about clear hypotheses, it’s also vital to communicate. You need your client to understand what could go wrong – you need to have their dev/UX teams on your side – and everyone needs to know what they are meant to do so that your experiment runs like clockwork.
Paddy’s top tools
– an A/B split testing tool that came up a lot during the day.
– a survey tool that also got a lot of love from the speakers.
The final stage: reviewing the data to see if your hypothesis was proved correct. You’ve got to be careful to ensure that you are checking for anomalies that could be affecting your results before declaring your experiment a success, and the key takeaway here is to be patient.
When you know you’ve got solid results roll out your changes (or not) and start again.
Cultural Conversions and International eCommerce by Joe Doveton(@GlobalMaxer)
Next up was Joe Doveton from GlobalMaxer, a company that has carried out conversion tests or cultural research in 23 countries. He started up by outlining the two approaches to international CRO:
Completely localised sites – you adapt the design, language and whole approach of the site depending on the target country.
Global template – all you change is the language.
Joe was very clearly in favour of creating completely localised sites and used a whole bunch of fascinating examples of cultural differences to demonstrate why that really is essential:
English use. Your international audience will have absorbed Engligh into their language, whether you know it or not. Looking for business shirts in Germany? Don’t ask for an arbeit hemden, you want a business hemden. Obviously.
So what does this mean for you? Use native keyword tools (and presumably native speakers too).
Context. Some cultures are high context, meaning they derive meaning of images/words/symbols from the context they appear in. Others are low context and can derive the meaning of a thing from the thing itself. (I’m sure Joe Doveton had a far more eloquent way of expressing that than me.)
The takeaway? You may need to completely rework your websites for high context cultures, like the Chinese, who have a high propensity for uncertainty avoidance. And following on from that, written is only 1 part of the mix.
Colour matters... but maybe not that much. Broad statements such as “red symbolises luck in China” may be true, but they only tell part of the story and they won’t necessarily equal success.
What you need to know is that there are no shortcuts to building the ideal website, you have to run interactive tests.
Remember that fashions change, do your technical research and don’t make any assumptions.
Relationship to money. Cultural attitudes to money may have a huge impact not only on conversion rates, but also on how a site’s payment system should function. Among other interesting stats came this:
Again, you should research your market and get to know your audience. The cultural attitudes to money may affect more than your checkout page.
But all this barely scratches the surface of all the tiny differences in cultures across the world – what’s best practice in your culture, may be another cultures worst. Remember to test your assumptions, and learn to embrace losing tests.
And if those stats alone weren’t enough to persuade you that the time to make the move to mobile is now, he backed this up with great examples. Plus.net
introduced a responsive site, resulting in 10 times more sales and 2 times more traffic. Your design doesn’t even have to be all that – check out this mobile site:
With the case for creating a mobile site firmly made, Stephen went through a list of tips for creating your own mobile sites:
- Start small and scale up. Why not test just one page? If it works, you’ve made the case to make more mobile pages.
- Create mobile/tablet/desktop dashboards in Google Analytics that focus on your goals for each device type.
- Create mobile-only surveys. (2nd recommendation of the day for Qualaroo!) Ask questions like – “what was the purpose of your visit?” “what nearly stopped you buying?” and really importantly, “why are you switching to desktop?”
- Use desktop click maps to help prioritise your site and pick the pages for testing.
- Dogfood your site on mobile, i.e. use your site like a customer would. It’ll help you to really understand what’s going wrong (or right) with it.
- Identify key mobile personas and determine why someone might use a mobile instead of a PC.
- Run usability tests with sites like usertesting.com. Don’t assume you can guess how people use your site. If you need convincing of that, see what happened when Google asked people “what is a browser?“, which anyone working in digital can explain.
- Sketch designs on business cards with sharpies*. It’s about the right size for a mobile screen. This is one of those “why didn’t I think of that?” genius tips.
- Create wireframes for your responsive sites. Stephen suggested using Axure or Adobe Edge Reflow.
- Two mobile sites to copy: Home Depot simple pages, focussed on localising traffic with store finder information prominent and pick up in store options. Target clear images, simple information and a basic shopping cart.
- … And one to avoid: Gap which overcomplicates things, asks for irrelevant details and frankly makes odd assumptions (the first two card type options are GapCard or Banana Republic card – especially annoying when card type can be determined from the card number alone.
- Small changes will help improve your site, but sometimes you need to do something totally different to get great results.
*I’m sure other pen types will suffice.
The Value of Social by Philip Sheldrake (@Sheldrake)
Philip Sheldrake had perhaps one of the more difficult topics to cover for the day, talking about how to prove the value of social. As you might expect, he didn’t give any conclusive answer, what he did was pose questions to help you determine what to measure.
His opening gambit was to say it’s wrong to talk about social in terms of ROI, because it isn’t all about selling. Selling’s just one aspect of marketing, and marketing can be defined as:
The process by which companies create value for customers and build strong relationships in order to capture value from customers in return.
And his opinion is that social is 100% focussed on value creation – and not on monetary gain – so to measure its success in terms of financial reward is nonsensical.
So where does that leave digital marketers that have to show value? Philip referenced AMEC’s Barcelona Principles
to discuss the issues:
- Everything about your business – from your reason for being to your marketing strategy – is unique, so the metrics that you choose to measure success have to be unique too.
- Outputs, e.g. likes, mean nothing, talk about outcomes.
- Your methods for measuring success have to be transparent, i.e. anyone should be able to use your method at any time and reach the same conclusions.
- Use a handful of complimentary metrics to measure performance – a single metric is too easy to game, which is damaging to everyone.
- Talk to business people in terms they understand.
Ultimately this presentation was more of a call to arms to marketers than a practical how to guide, which makes it challenging to describe. Maybe Philip’s free ebook, Attenzi, which he pimped out at the end of the talk will help us all to find some clear answers. Or maybe it’ll just send us further down the rabbit hole.
The guys from Decibel Insight delivered a talk that showed us their Analytics tool, alongside some pretty cool insights. They were focussed on the difference between Analytics – monitoring performance – vs Insight – using data to improve performance. And while they realise you have to report on what’s useful to stakeholders (KPIs focussed on traffic and growth) they suggest that it’s more useful to marketers to track content to understand engagement.
According to them, only 14% of businesses are tracking content, so to make it easier on all of us they gave us 6 steps to do just that:
- Learn about your visitors.
- Experience their experience.
- Find your most popular content.
- Ensure that visitors can see it.
- Learn how visitors navigate.
- Increase the effectiveness of mobile.
The most useful takeaway from this presentation was a simple, yet significant idea: 98% of visitors (on average) don’t convert and that skews all of your analysis. Segment your converting traffic for real insight into the needs and pain points for your customers. It’s obvious when you think about it really, but it affects everything: Any changes that you make to your design needs to be with customers in mind and without segmentation, you could inadvertently make your site less attractive to the people that matter.
Why do you hate Excel? by Russell McAthy, Stream:20 (@therustybear)
I’m glad to say I wasn’t the only one tweeting “I don’t!” in answer to this question. Yes, my name is Katie Walton* and I am a geek. With the support of my fellow geeks, I’ll get through it. Or judging by today, I’ll just get geekier still.. oh well.
The slides really do speak for themselves – Russell provided a bunch of great Excel tips to help everyone use it better in an adorably excitable geekfest. Read them, marvel at their wonder, then go put them into practice. Russell, you are now my hero.
*In case you wondered, yes, it was Katie Saxon, I got married last month 🙂
Attribution with Google Analytics by Dara Fitzgerrald (@darafitzgerald)
Dara’s talk was focussed on the Multi Channel Funnels (MCF) reporting that is available in Analytics. In case you are unfamiliar with this, it shows how each type of traffic contributes to conversions to give a better understanding of how your marketing is working as a whole. Here were Dara’s tips on using the tool:
- Conversion tracking – be it ecommerce tracking or goals.
- Proper campaign tagging so you know that you’re attributing the right results to the right campaign.
- Understand that Assist and Last Conversions are not mutually exclusive – i.e. organic traffic may appear twice in a conversion path, counting as both an assisted and a last click conversion. Use conversion segments (include assist, exclude last) to separate out true assists from self assists.
- As with nearly every speaker today, he talked about the need to customise the set up for you, yes it’s good to know that email marketing assisted conversions but was that a general newsletter or a triggered email reminding someone their subscription is coming to an end?
- Look at conversion paths to see where certain channels appear in the chain, and as a result, to understand how your strategies should work together.
The subtitle for Dara’s presentation was “How to do it … and why you shouldn’t”, he was quite clear on the downsides of the tool – and more specifically, the ways it’s used:
- Device proliferation means that the data isn’t accurate – the data doesn’t show cross-device paths, e.g. discovery on mobile, research on tablet and conversion on desktop, for example.
- The tool doesn’t consider the impact of offline on profits.
- Attribution is retrospective and often isn’t used to create hypotheses to test and improve marketing.
- … That said, MCF only shows what happened with converting traffic, so potentially you’re building tests on the basis of flawed data. (Unfortunately the solution to this issue is to go premium, with the hefty price tag of £150,000.00)
- The different attribution models can be used to make people look better at their jobs.
- Attribution focusses on acquisition only – this doesn’t tell you how to build lifetime value. You should apply your testing to your retention strategy too.
All of that sounds like a great case against using MCF, but really Dara’s just preaching caution. His key takeaway was that not having a perfect situation is not a reason not to do anything. In other words, maybe you can’t draw perfect conclusions from the tool, but you can create interesting hypotheses to test.
Getting Google Analytics Going from Get Go by Nikki Rae (@AnalyticsGirl)
Nikki Rae’s talk was a a run down of why your Analytics is lying to you – and what you can do to fix it. Featuring such problems/solutions as:
- Your referral traffic includes organic search engines – Analytics only registers a handful of search engines – add code to your page to force it to show the traffic as organic.
- You have traffic from keywords that you don’t rank for – your code is on someone else’s site – use a filter to weed out the incorrect traffic.
- You have multiple versions of pages in your Analytics reports, e.g. / /home.asp /Home.asp – use a find and replace filter to create aggregate stats.
- You don’t get organic keyword data anymore – set up internal site search for access to a different kind of keyword data.
Her slides act as a great checklist for ensuring that all is right with your Analytics set up – bookmark for future reference!
The Power of Segmentation in Web Analytics by Anna Lewis (@Koozai_Anna)
Everyone who comes to your website comes for a different purpose, you need to look at the detail to really be able to make sensible decisions.
If you can segment your audience, you can give everyone what they need – and that helps you to get more results from them. When you figure out the segment that is most valuable to you, you can target them to get better results in the future.
As with most other speakers, Anna made it clear that there is no one size fits all solution. While she gave examples of ways to segment audiences you have to figure out which segments matter to you.
Real Time Business Insights by Ryan Gallagher (@ryangallghr)
I’m really sad that these slides aren’t available online – Ryan delivered some cool facts and great pictures, even though there were few practical tips. It’s hard to really give the gist of the talk, but the key theme is that data without context is meaningless.
Realistically the average company doesn’t have the money to look at big data effectively, but you can start overlaying information from the real world against your Analytics data to improve your performance. This will be easier when Universal Analytics launches, but in the meantime, segment your data and compare with other data that you have access to. Use your brain to join the dots and gather insights – it really is the best big data tool
Customer Value Optimisation by Andrew Hood (@lynchpin)
Andrew flipped the usual idea of acquisition vs retention on its head, arguing that retention is a short term strategy to stop something bad happening, whereas acquisition is about growth, which is inherently long term thinking.
Semantics aside, Andrew suggested that we should be using Analytics to not only suggest CRO improvements, but also to find patterns that can help improve the value of the customers that we acquire in the first place:
- Can you look at particular products that do/do not indicate quality?
- Can you identify form data that indicates quality leads?
- What are the key signals that suggest a customer will leave?
And the key focus of asking all of these questions is to determine whether you can then focus your marketing efforts and your CRO success criteria on customers that are more likely to have a high lifetime value.
What is Possible with Call Tracking by Ali White (@AlistairWhite)
Even in an online world, phone calls are hugely important to all businesses, as Ali showed with these stats:
According to Google, in the US, 70% of mobile searchers call business directly from its result pages.
71% of people expect support in 5 minutes or less.
61% want support over the phone.
Using call tracking helps you not only with attribution, but also with CRM. You can get super detailed with your call tracking, using separate numbers for each different campaign, whether on- or off-line. You can set up different types of numbers for desktop users (e.g. 0800) and mobile users (e.g. 03). There’s so much you can do, but you don’t have to so it all, keep your tracking top level if that works best for you, because ultimately you need to improve your business.
Ali’s overall takeaway: Don’t optimise for conversions or even for revenue, focus on profit!
Dealing with Not Provided by Dr. David Sewell (@seoeditors)
This was a lightning – e.g. brief! – talk introducing a (not provided) tool that uses available evidence, e.g. location, landing pages, browser to determine which organic keywords drove traffic to your site.
Thanks to Guiomar Gonzalez for sharing this picture.
Head over to www.notprovidedtool.com to take part in the beta tests for this tool.
Leveraging Your Competitors’ CRO Efforts by Yousaf Sekander (@ysekand)
And if that last tool didn’t do anything for you, this next one (almost) certainly will – cromonitor.com
. We all know that competitor research is invaluable for
finding ideas for links, ad copy, content and more, so why don’t we do it for our CRO efforts? Thanks to this tool you can record when your competitors are running A/B tests – and better still, you can figure out which test won. Awesome, right?
Sign up for the tool over at cromonitor.com.
As you can tell, it was a pretty packed day! Here’s to next year!