We’ve already covered Mobile SEO in our ask the experts series, and now we move on to CRO (conversion rate optimisation).
We asked our chosen experts the following questions. Click on one if you want to jump straight to it.
- Is it fair to say CRO is not as en vogue as it has been? If so, why?
- Is there received wisdom for what metrics conversion specialists should be measuring? And over what period?
- Mobile seems to have increased the checkout abandonment rate. What criteria/features are particularly important when creating ‘flow’ in the checkout?
- What are your absolute UX no-nos? (e.g. carousel, type of nav, type of content etc.)
- What are the best tools for CRO? Has Google Optimize been widely adopted?
- What sorts of dynamic content and strategies are cutting-edge optimisers experimenting with?
- What are the pros and cons of machine learning solutions?
Paul Rouke, founder & CEO, PRWD:
Yes. There is a divide taking place in the industry. The businesses at the strategic and transformative level of maturity don’t even use the term CRO or conversion optimisation. Experimentation is the mindset change they have adopted and made part of their DNA. Brands like Amazon, Skyscanner, Booking.com, Shop Direct, Airbnb all have an “experimentation culture”. The world’s largest testing tools now position themselves as Experience Experimentation platforms.
Where does that leave the rest of the industry? There is continued mis-understanding and under-appreciation of how conversion optimisation can be a catalyst for a business to become customer centric. In 2017 and beyond, brands have so many “sexy” areas for potential marketing investment, such as social media, programmatic, AI, big data and personalisation, so CRO in its basic sense is seen as tactical bolt-on.
Did I also mention that over 95% of marketing spend is still invested in acquisition? This is a shocking and worrying reality.
Airbnb has an experimentation culture
Depesh Mandalia, CEO, SMCommerce:
CRO is still pretty important, however my observation from SMEs is that they’re unwilling to commit to a single person overseeing CRO, or at least remain unconvinced that CRO is in itself a team or person. Instead more and more marketers are adding the CRO string to their bow, creating landing pages, running split tests and analysing results.
The risk here is that the deep technical and statistical expertise that a CRO brings is often overlooked, with credibility given to tests that on the surface look great but lack impact when for example moving from 50% of traffic to 100%, because of statistical significance. There is more a CRO person can bring, but justification is that much more difficult.
Tasin Reza, CRO Director, Redeye:
I wouldn’t say it’s no longer in fashion. I would say it’s now becoming part of the whole success strategy. We are noticing that the companies are now getting their internal CRO team to focus on growth.
Greg Randall, MD, Comma Consulting:
Any organisation that offers CRO as a solution is a dinosaur. This is an old outdated approach and philosophy in improving the performance of the digital channel. The view of conversion rate optimisation is to try and make more money and improve conversion rates for the retailer. While this is important, it forces the business down an internalised view of success and puts the priorities of the brand/retailer first when in fact it should be the customer.
CRO brings with it old outdated approaches to improve performance. Many CRO practitioners drive this function via AB testing. While this is part of a successful recipe, it is a very small piece to a larger puzzle. If AB testing is conducted on pages where digital best practice is not in place, the results will not add value. It’s like having a poor control group in the context of science.
So what should take CRO’s place? Retailers need to replace CRO’s with “CCEO”’s…. “Chief Customer Empathy Officers”. A CCEO focuses on creating amazing online experiences and comes from the perspective of the consumer. Through this focus, consumers who have amazing experiences with a brand will be far more likely to purchase and develop loyalty.
CRO as dinosaur?
2. Is there received wisdom for what metrics conversion specialists should be measuring? And over what period?
The first thing to make clear is that “conversion rate” should never be looked at as the one main metric. Optimisers should be looking at broader business metrics, AOV, retention, RPV as well as ROI from their largest acquisition channels.
Experimentation efforts should be measured both in the test period but also following the implementaion of successful tests. Post implemention testing, to understand the reality of how the change is actually impacting user behaviour and business metrics, is hardly ever done.
When measuring performance, it’s far more intelligent to look at funnel performance and other key indicators such as the percentage of visitors who are being sent further through your discovery and purchase funnel – there are simply too many external factors influencing site-wide conversion to make this the primary success measure – it’s just a shame that the industry acronym is CRO!
For the seasoned CRO practitioner, the measurement goals are quite clear; at KPI level, macro level and micro. You can’t shift the click-to-buy conversion rate through a single test – instead you can impact a macro or micro conversion along the customer journey.
Even then segmentation, for example by marketing channel, further muddies the waters when you consider how buyer behaviour will differ across channel and thus a low conversion rate on a top of the funnel keyword search may not simply be a problem with the website, but a natural part of the journey for a particular person. Therefore it continues to be of upmost importance to consider buying cycles when deciding how long a test should run, not simply to wait for statistical significance which does not take human behaviour into account.
A typical customer journey can be complex
Matt Lacey, performance director, Code Computerlove:
We’re swimming in data these days, but a clear measurement strategy and well-defined KPIs are all too often lacking. Metrics will vary significantly depending on business model and business life stage.
Defining 3-4 actionable metrics to focus on, and measuring them with consistency, is a great way to start. Fundamentally, you need to understand how metrics are related, which ones demonstrate success, and which you can influence.
For ecommerce businesses, global conversion rate is a default metric, but revenue per visitor may be a better metric that helps you understand conversion, abandonment and AOV in one figure.
The Chief Company Empathy Officer works beyond behavioural data to gather value-driven insights to broaden and strengthen amazing retail experiences. The mix of data sources required are both quantitative and qualitative.
Broader and more reliable data sources include:
- Behavioural site data
- Customer Support team data
- Chat data
- Employee interviews (interviewing those on the coalface to identify consumer pain points)
The insights gathered from support teams, chat, and employees deliver context to the behavioural data and enriches the insights from the consumer’s perspective. This clearer view of consumer buying/information gathering journeys enables the Chief Company Empathy Officer to provide more value-driven changes to both the online and in-store experiences.
3. Mobile seems to have increased the checkout abandonment rate. What criteria/features are particularly important when creating ‘flow’ in the checkout?
It’s all about simplicity and pure usability – remove unnecessary fields, use large, finger-friendly form elements and buttons, dynamically change the keyboard based on the input, chunk up fields so as not to overwhelm visitors.
User research is one of the most effective ways to quickly understand how usable your mobile checkout is. I recommend brands follow in Schuh’s shoes (so to speak!) – don’t accept that “mobile conversion rates will always be less than desktop”.
Providing Paypal as a primary payment option also means visitors don’t need to mess about with their credit card.
Mobile has not seen a significant shift in conversion rate over the last few years, however the same report signifies an increase in multi-device attribution which could suggest the focus on decreasing mobile dropout could be on enabling a better multi-device experience.
For example if you have a fairly complex product that needs configuration and then consideration, offer the user the ability to save their configured product to email to view later, most likely on desktop or tablet perhaps if they want a better look later.
I always say Keep Things Simple Stupid (KISS) and for mobile, this is even more important. The fewer steps users have to complete on mobile, the better their experience will be. For example, we have seen an increase in mobile conversion rate with the introduction of the PayPal express checkout.
Firstly, are you conducting user research? If not, then how can you start to know 100% that some of your UX isn’t working for your visitors? If and when you develop an experimentation culture, nothing is specifically a no-no – you will research and potentially test what is working or not working for your visitors.
Lots of research has been put in on the use and impact of carousels and this continues to remain a UX no-no. The primary problem is of user focus and the impact on cognitive load. The more the user has to think about or decide between, the less likely they are to make a positive choice.
Another definite no is music auto-play, which Facebook recently announced they’d be enabling by default for videos. There’s nothing that annoys customers more (perhaps) than visiting a website with sounds on auto-play when you’re not expecting it. This happens on some news and media sites which run auto-play videos.
Nothing is sacred when designing an experience, but there are design principles and common heuristics that provide solid foundations for a great experience.
We have seen examples on different sites where features and functionality perform quite differently based on context and environment. For example, carousels have won in some A/B tests and lost in others. Both long-form and short form landing pages have won in different contexts as well.
If we can move more quickly through cycles of good design using common sense conventions, conducting user research and split testing, we will ultimately build better experiences.
There are a couple of things:
- Don’t try to trick the user into doing something. While you can get a short term gain, in the longer term this will have a negative impact on the brand.
- Don’t overload the user with too much content at once. As humans, our attention span is reducing and based on a study I came across, it’s less than a goldfish! What this means is if you try to give a lot of unnecessary content that’s not relevant, users will struggle to find the content they are truly after.
Based on my personal experience of working with a variety of tools, Optimizely is still the leading tech for CRO. Google Optimize is still in its infancy from a testing tools’ perspective. However, for companies who are just starting their CRO programme, this can be a great tool to prove the return on investment of CRO.
The most important tool for businesses is what they have huge access to – their employees’ brains. So often people and their expertise and ideas are overlooked or under-utililised – or simply not given the time to apply themselves to conversion optimisation strategy.
Tools and technology in the classic sense are purely the facilitator to run experiments. There are too many companies who have “all the gear, and no idea” in my years of experience. There are also huge levels of investment made in tools that simply aren’t being utilised to anywhere near their potential.
As a marketing agency, we work with clients solely on Facebook advertising. However we also have a set of indispensable tools we convince clients to adopt from Hotjar, Google Analytics, a landing page tool like Leadpages or Unbounce and an email automation tool like Autopilot HQ.
For any marketing channel to be effective, each part of the journey needs to be optimised from the first ad impression onwards. CRO therefore becomes an integral part of the marketing work we undertake to enable us to take joint responsibility with our clients on converting prospects.
Businesses that have embedded an experimentation culture are now focusing a lot of attention on behavioural targeting, advanced personalisation and programmatic.
I believe the best thing about CRO is the opportunity to be bold, be innovative and have a fail-fast approach. We have been doing some amazing experiments to try out things that no one has tried before. This allows us to test and learn and move forward with bigger wins.
Whilst old school for some, we’re still seeing the benefits of dynamic landing pages based on advertising content. For example sending a link from a Facebook ad into a landing page, with data in the URL which the landing page interprets allows us to serve content and creative on the landing page to match the ad.
It may sound trivial yet being able to dynamically create landing page experiences on the fly allows us to scale faster.
I’m really excited about some of the areas where machine learning is being applied to optimising experiences. For example, using machine learning to identify segments that are under- or over-performing based on a wide range of factors, and then tailoring their experience to improve performance.
However, there are a few caveats to consider. Machine learning can help us to optimise within a fixed set of constraints, but this leads to a set of limitations. How does it account for brand or lifetime value? How do we identify the size of opportunity within those constraints? How does it impact our approach to broader innovation to grow out of local maxima?
The challenge is how to balance human judgment and machine optimisation, but get this right and the opportunities look exciting to say the least.
I attended Turing Festival last summer in which Oli Gardner (co-founder at Unbounce) gave an amazing talk on how they’re using machine learning to optimise landing pages before they’re even launched to real users. The fascinating part is that what they were building would become self-learning. It was still in the works and who knows how long it would take to mature the technology but a company like Unbounce, with the millions of data points they have, could really help accelerate learnings for any company, but especially those that suffer from low traffic.
Pros – initially heavy resource involvement but later it would automatically provide the best experience. Cons – lack of control.