Interview with Oliver Paton about Conversion Conference

Oliver Paton

The next speaker to be interviewed as part of the series leading up to Conversion Conference London 2013 is Oliver Paton who was a Senior Web Analyst with Hotels.com. Oliver has worked in digital media for over 7 years in both agency and client environments and grew Hotels.com testing program from 5 tests a month to over 20 tests a month.  He is now transitioning his role with Hotels.com to a web analytics consultant with the Accenture Digital team.

Oliver will be presenting his Conversion Conference session “Optimizing Your Optimization” during track 2 on day 1 from 2pm.

  1. Is business ready for the mass of data that universal analytics produces when exploring what could be optimised/tested further? Universal analytics is the first step towards what is the Holy Grail in optimisation testing… The possibility of targeting users across multiple devices and channels, and delivering the same treatments to users across these devices and channels. Being able to measure across these channels and devices is the first step. Then the ability to stitch data in real time in order to segment users to serve experiences that are relevant to them. I don’t think many companies have perfected this, mainly because in most cases it requires the user to be identified in a unique manner. In most cases this requires them to be logged in.
  2. Also Is real-time conversion optimisation always possible as part of continuous revenue improvement? Real time conversion optimisation is always possible, but depends on what testing solution you are using. In some cases it may require development work in order to take final solutions into code. In some cases if an optimisation idea is simple, it can be done in the testing tool, but some features that require back end development will always be dependant on development teams and hence developement backlog. When a company is completely engaged in optimisation practices, this becoes a lot easier. There has to eb a focus on making sure that developement teams are working in a way that allows optimisation teams to be involved in the work flow processes.
  3. Do you feel that analytics platforms are getting better for tracking & analysis of offline conversions? I think this is purely down to the implementation. If it is set up correctly, then it is always possible to track offline conversions. As analytics packages get better, then it makes it easier to implement these solutions as opposed to having to find a hacky way of getting the data inot the package. If you are passing all your Web analytics data into a database, then it is likely that the offline orders will be there anyway, making the value of having those conversions in the web analytics package fairly low value when the data is all in one place already. I am a strong believer in web analytics data being passed into a database anyway, as it is valuable for BI, and more useful in that format.
  4. Do you think people rely on attribution modelling in analytics software too much now to develop their own? I think that ad server attribution modelling and analytics attribution modelling are improving to the point that investing in development of a new tool might be going a bit far. I think that the real value will be when people start conducting fair tests on their models.
  5. How do you think the US/Asia sees the EU’s strict data protection directive? Will they follow the EU’s lead? I would hope that they don’t follow suit. Advertising and the ability to track users has made the internet what it is to this day. Advertising paid for the internet. I understand that companies should respect people’s privacy, but the data protection directive, and the hype around it has the potential to create paranoia in the general public when in reality MOST data collected is anonymous and used at an aggregate level.
  6. It’s been a year since the EU cookie law, do you think business adapted well to the change? I think that some companies have taken it to heart and done what has been asked. I have also seen companies simply ignore the directive in a “see what happens if we do nothing” approach. I have also tested a solution in an A/B test and seen negative conversion as a result of notifying users that cookies are used. I wonder how the IAB and all other governing bodies feel about companies losing out on revenue to implement these solutions.
  7. How do you deal with a global platform and conversion optimisation with multiple languages? In many cases I have tested in English only, and then if the test is a winner, then it is moved out to other languages. In tests test’s that have to work everywhere I tend to refer to a lookup table in order to find the translation of the copy to be changed. Sitespect are working on the ability to upload tables into the system in order to do exactly that. In other JS based tools, it has to be coded into the page.
  8. When do you know it’s time to let go of a proposed test when you get pushback? Interesting question! I will be covering this in my presentation. I think it is time to let go where there is a justifiable reason for the test not to happen. This generally only tends to be when the test has been run before. This is where tracking or past test is highly important. Creating a test results library is highly valuable.
  9. What stakeholders do you need to get onside early in a optimisation project? I would tend to say, start at the top. Engagement from senior management, and communication from then about optimisation is highly valuable. This is sometimes tricky as they can the quite far away from the coal face when it comes to optimisation. After this, it is always good to have some developers in your back pocket to help complete the build of test variants.
  10. Do you consider the impact on lifetime value of a customer when evaluating conversion optimisation? Only if we suspect that it may have an effect. When I was working for hotels.com we were testing a compare tool feature. When turning it off, we knew that some users might have items saved in their tool. In this case we would only stop new users from being exposed to this variant, and users who had been assigned to the test variant in the test period would remain seeing the compare tool. In other cases we have looked back at test to review latent conversions for users that have come back and purchased. In most cases I have not seen a difference. Going into the longer term, I feel that something that improves conversion in the short term is probably unlikely to affect customer loyalty more than simple after sales care.
  11. What is one vertical that you think always leads conversion optimisation? After working for Expedia for 3 years, and having spent a lot of time looking at competitor’s, I would say that travel do very well, although I know that gambling sites also test a lot. If we looked at a correlation between number of tests and online revenue it would correlate quite strongly…. as it should. Testing is really a due diligence in the modern website development model. With so much money invested, and so much money at stake, it seems dangerous to leave the front end untested. Also, increases in conversion help justify more budget on web development and paid media to drive traffic.
  12. What’s some new software/tools that you think people should consider trialing or exploring? Sitespect. In my opinion Sitespect offers so much flexibility for your site, and allows you to change anything and everything on the fly. We have been in a product governance meeting at lunch, and then had a test running by 1pm based on the ideas expressed in that meeting. That’s under an hour from blue sky thought to test start. Because it uses regex, it is also very web analyst friendly.
  13. What is the main points that you want people to get from your presentation? I am hoping that people will take a step back from their optimisation processes and look at how they operate, and recognise how they can improve the process and work better with the rest of the teams within their organisation to get more from their conversion optimisation program.
  14. What are some of the other sessions/speakers at Conversion Conference that you are looking to see? Are there any other Data Driven Business Week conferences or sessions you are looking forward to seeing? Paul Rourke was a big influence last year on our optimisation programme at hotels.com. This was partially because he was talking about booking.com, but also because he influenced us into thinking about conversion tests by theme in terms of what concept of influence the test used. This helped us prioritise tests and make sure that successful test themes are tried out across the site as opposed to in just one place.
  15. Where can people find and engage with you online? Linkedin and @Ollypaton

Thank you Oliver for your time in answering these questions.  If you want to catch Oliver speak and save ££££ on tickets for Conversion Conference London you can use our discount code LOSTAGENCY13 and you can register and find out more here.