My life as a Web Analyst – February 2011

This is another post on my hands on life as a web analyst. Week 8 was a busy week. Lots of things happened, good and bad. Although we are getting to exciting times, a lot of work is being released so we will see the effects very soon.

  • The week started with looking at the numbers. Both at the customized dashboards in Omniture and some ad hooch analysis of the effect of a landing page. Unfortunately the data was not enough to draw any conclusions.
  • The dashboards that I have been perfecting for a month or so is proving to be up to the task. I spotted a small trend break and reported this to the stakeholders. Small wins are always good.
  • I had a meeting with the designer discussing wireframes on a few pages. I laid the ground and got some help to finalize them. Unfortunately we are lacking an Interactive Designer so I have to be hands on with wireframes, which is a bit outside my area of expertise – notice the understatement.
  • A colleague of mine wanted some help understanding the behavior of some key pages. We investigated why people did not convert and the answer was very clear. Immediate improvements of the pages had to be made.
  • Meetings with two of our tech groups. Both of the groups have new members so I have to have several meetings with them to set the long term agenda on where we want to take the sites this year. I find it very important especially in the beginning to have everyone aligned towards the same goal. It saves a lot of energy in the long run.
  • Together with the designer we “invented” a new landing page for a financial product. Let us now see how it performs. To be honest the one we have today is very ugly but this website has blown my mind. Everything we throw at it, it rejects more or less. Conversion increase is very hard I have found out.
  • The bad news of the week was the new Google algorithm update. It punished some of our sites, making people almost get into the panic mode.
  • Finally this week we have prepared for the redesign release next week. We have reported bugs and made quality checks all week long. Let’s see how it will be received by the users’

My life as a Web Analyst – January 2011

When I meet new people they ask me what I work with…and to be honest…explaining is actuall not that easy, especially for people outside the online sphere. This gave me an idea, to write small posts recaping the week that has pased explaining on some of the issues I work on. I this way you might better understand what a conversion specialista aka web analyst job is truly like. I promise you  it is not as glamourous as you think…

Week 2, January 2011.

  • First week on the New Year. We had hoped to have a new site release up by now but now it looks it will happen in the end of the month. You know how development is, not a straight road.
  • I documented on the whiteboard the development process for the rest of the year. This has to do with the previous post I wrote about conversion increase as part of the business development process. What I have done is to put small notes on the whiteboard to visualize the actions we will work during the year.
  • I have worked on doing the same thing for a second website. The progress of this website is not as far as the first one mostly due to the fact that a vision of the business is lacking. I cannot stress the importance of having a vision that is driving you.
  • We had several meetings with our designer concerning the new logo. We finally decided that we want the logo to feel more newsworthy than corporatate. The logo that won the logo contest was not used at all – waste of money but it assisted to narrow our current view.
  • I had an interesting meeting with one of the guys working with the core analytics team. We discussed what we could do with all data we have collected. We share the common view that the difference between companies that become sucessful in todays online environment are the ones that use data in their decision process. Personally I want us to move towards a predictive analytics path and ultimately towards prescriptive analytics.
  • I created a handout document describing why and how one can increase the conversion for casino websites. We designed a new casino landing page according the recommendations. I must say it looks good lets now see how it performes.

How I work to increase conversion

On my current position a lot of people rely on me to increase conversion on their sites. Increasing conversion are in many cases not as straight forward one many think. If you have worked with it you might testify on the frustrating side of trying ideas that do not show any lift. On the other side I see a lot of successful cases of button placements that can increase conversion but those changes on a website or a page are few and you have to be a bit lucky to find them straight on. What many people don’t tell you is that when you have one successful test you most likely already have done a few unsuccessful tests.

My method for increasing onsite conversion is to work hand in hand with the business development of the site. My view is that sustainable online conversion can only be achieved when you understand the need of your users, and then applying this knowledge to the development process. Unfortunately this view is not commonly shared among online companies since many of them started small, grew and now is in a phase when growth is all about user knowledge tied to the business development – a phase that requires different skills than when you are a startup.

The other day I picked up this book and boy I was glad when my view was confirmed by the authors (The book is called: “Conversion Optimization; The art and science of converting prospects to customers”). This is the passage that elaborates on my view:

“The question is not whether you can achieve a double-digit conversion rate; the question is whether you are willing to do what it takes to achieve a double-digit conversion rate.

Expectations for conversion optimization results should be combined with realistic expectations regarding the amount of work required to achieve these results. Because many clients do not understand the investment and commitment required from their team, they give up too soon. Conversion rates won´t increase overnight. You must start by understanding the visitors you are trying to convert. You then have to create hypotheses about these visitors and why they interact the way they do with your website. Then you have to validate these hypotheses by making changes to your website and tracking the customer response. All of these steps require time resources and a financial commitment, which is why they must be done accurately and with great focus.”

So two keywords can be extracted, commitment and knowledge of the user intention. When you do a new conversion test you create a hypothesis, design your test, run it and confirm or deny your hypothesis. Thereafter you probably need to refine it and test again. This cycle might repeat itself a few times.

So far this process has worked well, with conversion increases of up to 40% but lately I am facing more and more obstacles, since the low hanging fruits have already been picked. Now I am trying to steer more and more towards the business development process adds the complexity since you also have to manage the expectations of various stake holders – a fun process but nontheless not as easy.

My advice to you is to inform your stake holders that quick results are actually not that quick in reality. Knowing what to test and doing it successfully in the long run demands a fair amount of deep user research and strong comittment. There is no way you can cheat on the users investigation step, since it is your fountain for testing. Remember to drag your stake holders into this since you need their long term comittment to be successful.

Good luck, and if you have questions dont hesitate to ask…

Why I think Amazon can increase their conversion

The other day I tried to make my first Amazon purchase but failed. The process of ordering was nothing but smooth and straight. I even struggled to figure out the price of the product I wanted which makes me wonder, am I a total weirdo? If I that have extensive Internet experience and have setup my own webshop, worked with usability and seen the most on the web have problem ordering then how many others are not in the same position with less experience. Amazon is one of the highest converting shopping sites out there making tons of cash but I wonder if this is not due to their strong brand recognition and other services.

The other day I heard they are selling 73 items per second and they have a conversion rate somewhere between 5-7% which is extremely high for an e-retailer which usually is around the 2% mark. Amazon though is really good at expanding their business model into areas that they will make money. For example today they announced that they would start selling movies on demand competing directly with Netflix.

To be honest I have always wondered why their site looks like 1999 in a time where we have extensive tools for web analytics, usability and AB testing. I know they have a good web analytics team and they do a lot of testing but from my point of view their pages are so cluttered that as a user I have big problems understanding what I should do next, which in all known cases is a conversion killer.

What I think is wrong and what they should do to change it

On the product page first of all I had a hard time finding the price. It is cluttered around text which is trying to convince you by reinforcing the good things. The add to basket button is far away from the price. I would like to see the price much clearer and closer to add to the basket button.

The up selling part on the page is a very clever element. I guess they get a bit more money from this.

Clicking the next button I get a page which trys uppselling again, athough the page is very cluttered and the next step is not that clear.

The payment page is a bit clearer but still not as clear s it could be.

All in all I think there could be some improvments but the question is why have they not done it? Or have they and not managed to improve conversion. All my experience tells me though that a better usability will increase conversion.


Why have they not tried a better usability/design?

Here are my my five cents on the above question:

  • Why change a model that works. The problem with this is that you don’t know your conversion roof, how high can it be?
  • Changing one page design might require you to change all other pages.
  • This might sound silly but the cluttered pages might make up selling easier.
  • Internal politics.

What they are world class in

So, all in all I think Amazon most likely knows their stuff very well but the issues I raised above are more or less best practice according to my opinion. Having that said let me give you my opinion on what they are world class at:

  • Up selling. In several steps they try to up sell items.
  • Understanding what users are looking at and getting more of that i.e. user reviews.
  • Behavioral targeting – displaying products based on your history.
  • Getting promotions to the right people.

How we increased the conversion by 40% on a casino landing page

I want to share with you one of our successful test we conducted in 2010 on the online casino market. I want to do this because I want to give something to the online gambling community, a community that as I have experienced it prefers to live in the dark and not share that much between each other. I also think that we in the casino market have to start listening to the customers much more and try to build sites that talk to the user needs.

If you are familiar with our company you will know that we have worked extensively with improving the conversion on several gambling sites both within casino and poker. To our disposal we have used almost all research methods that are available such as on site surveys, ab tests, external surveys and focus groups and of course web analytics. This research has culminated to the the creation of three profiles/personas which has been our guiding star to increase the conversion on a casino landing page.

How we started

We know that all traffic goes through one specific page, the landing page as we call it, which made it a logical for us to start at that point. After clearing the technical hurdles to do ab-testing we started by designing the first landing page without having the personas defined yet. Of course this was a failure. The conversion did not increase at all, on the contrary, it decreased. So we had to go back to the drawing table. Test number two did not show either a positive lift so we went back and asked ourselves why? By this time we had defined our three personas.

  1. The first persona  was the bonus junkies, people who are totally focused on the bonus and do not care much about anything else.
  2. The second persona was the game focused player. The one that liked a specific game and wanted to play just that.
  3. The third persona was the more general type that is just looking for a casino to play at.

We took these three personas and applied them to the design.

This is the original design…

…and this was the winning design where you have clickable tabs for each game type…

The picture below you can see how we thought when we applied the various personas to the design. Notice how the general feeling of the design is to convey a professionalism and seriosity something that is important for all profiles but a bit more important for the third persona type.

The result +40%

From the very start the new design outperformed the original by 40% in terms of depositing players. We could see that the tab functions was very popular.

By now you will probably ask yourself, will this work for all online casinos? Well to be honest I think so. My two strongest reasons are that:

  1. We see more or less the same user behavior on all casino sites we have worked with.
  2. Multiple independent surveys point in the same direction when it comes to the casino user profiles.

Various people within the web analytics business have preached about the importance of segmentation so I will not go into that discussion, but this test shows exactly how important segmentation is. Finding your segments is like unlocking the potential of your website. For you who have seen the movie Matrix it is like the character who plays the keyholder.

When does it not work?

I think the design itself will work independent where your traffic comes from. Of course the composition of the traffic and the user intention will determine how successful this design will be. Looking around at today´s landing pages I can say that I am not impressed at all. There are a few low hanging fruits there and the biggest of them probably the above segments.

My view of the Conversion Conference & Emetrics in San Jose, USA, 2010

I had the luck to attend the first conversion conference held in San Jose, USA, during the time 3-6 of May, 2010. The conference was in conjunction with Emetrics that has been around for a while so there were plenty of tracks to attend. I think in total there were somewhere over 100 tracks!

The theme of the conversion conference was to explore various aspects of website conversion such as videos, ab testing, copy writing, usability etc. The Emetrics part is more about web analytics and less on the conversion part but of course they overlapped a bit.

I and my colleagues Rickard and Jenny traveled all the way from small Stockholm to San Jose, the capital of Silicon Valley to see and hear the most prominent people within the web analytics field (a 16 hour trip…very tiresome). Unfortunately you have to travel that far if you want to listen to these people at the same conference. You rarely get this chance in Europe.

The attendance on the conference came all over the world. There were people from Germany, Israel, South Africa, Poland and another Swede.

The feeling I got was that most people there were some sort of consultants. I think less than 50% of the attendants were from end user companies. I was also surprised to meet so few big company representatives. Maybe they don’t do web analytics or they already know everything.

So, if I do a short recap of the conference here are my high and low points.

The high points

  • Many tracks. A lot to listen to. As usual when you have such great variety the quality varies a lot but it was good to be able to choose what you want to listen to.
  • The venue. It was held in The Fairmont Hotel, a four star hotel with large spaces and large rooms for such a conference. Unfortunately some of the rooms were small and got filled quickly. Also the air condition killed me some times.
  • It was well organized. Everything run smoothly from the various sessions to the lunch breaks run smoothly.
  • Famous speakers, top of the line. Some of the most famous people within the field was there as speakers. For example the much famous Mr. Avinash was there.
  • Probably around 500 people or more attending both conferences. It was nice to see such large turnout, which indicates that people are getting more and more conscious of what web analytics can do for them.
  • Networking. Like all conferences you get a chance to network, and I got to listen to various peoples “problems” which was interesting. I recognized many of the problems – like handling a vast amount of keywords or doing AB testing.
  • My top 3 sessions.
    • The presentation on how videos increase conversion if you do it right. From my own experience I know that videos can increase conversion but the two guys presenting was humorous and very down to earth. They also showed some interesting examples that can be used in various cases.
    • The session about sharing knowledge within the organization was interesting to listen to.
    • The Expedia presentation. Expedia must most likely have to most advanced web analytics and conversion program out there. They use software and manpower to maximize their efforts in various areas. It took them three hard years to get where they are but I bet they got their investment back with interest.

The low points

  • Over half of the tracks were too basic from my point of view. Of course various people experience this type of conference differently but I think many of the tracks were to shallow for my taste.
  • Too many consultants presenting and they tend to have a tone of voice that tells you that their findings is the rule. Especially the more famous ones were more or less preaching which I think if you are a bit experienced within the field you see through it.
  • Too few company presentations. I like company case presentations. I think they give you a good idea on how you can solve your own problems while watching the likes of others. Unfortunately there were too few company case presentations. I would like to see much more but I guess they are hard to find.
  • Many of the presented cases by the consultants did not hold to scrutiny. On questions from the audience many of the cases had not been implemented and were rather a suggestion  from the consultant to their client, so the outcome had not been documented.
  • I am not sure if this is a bad thing but the knowledge on conversion among the audience seem to be very low judging by the questions and the twitter chat.


I think this was the best conference I have attended so far. Would I go again? Yes, no doubt about it. That it was held in Silicon Valley was very interesting, I got the wibe that if you work with IT business development this is where you should be.

I keep asking myself how far along are we at Web Guide Partner on conversion and web analytics issues? Well, I truly think that we are among the most advanced companies out there but not on everything. We have explored most areas and have a broad spectrum of knowledge but we lack the marketing part such as ppc and offline advertising. Also our experience with social media is halting. The vast amount of websites we work with also limits us to the amount of deep diving we can do.

All in all though I am glad I attended the conference, it makes me a smarter web analytics and conversion guy, hopefully we can use some of the knowledge on our own websites.