We are looking for a data analyst! Check the job posting.

On April 18, our colleagues attended Emerce Conversion & Analytics at the Pakhuis de Zwijger in Amsterdam. One of the favorite events among marketers, analysts and CRO specialists. It was again a busy program with many interesting presentations. This year colleagues Ruben and Desiree were also on stage as moderators. The event was opened by none other than one of the founders of Online Dialogue, Ton Wesseling. Wondering what you missed? In this blog I share my favorite speakers during Emerce Conversion & Analytics 2024.  

Fast learning in seven weeks: the health insurance campaign at ING 

Tom Schreurs - ING 

There are only seven weeks at the end of the year when it all has to happen for health insurers. Then the premiums are announced and the period in which customers switch health insurers begins. Even at ING, this is the time when the entire team peaks. During this session, Tom shares valuable insights from three years of campaigns around health insurance at ING. He shares insights from interesting tests and gives concrete examples. 

Some context: every year 6-7% of the Dutch people switch health insurance companies. There are two peak moments that recur every year: the moment the premiums are announced and the end of the year because this is the last moment you can still switch. 

There is in this prime-time period of health insurance campaigns no time to waste When it comes to testing. Therefore, this is done in 3 stages: 

  1. Isolate the most promising hypotheses 
  2. Continued testing with these hypotheses
  3. The winner goes live and small UX things can still be tested 

There are a few more things important here. First, you have to be careful about generalizing generations and isolating by age. In addition, once you have released the app target groups, it is important that you start testing based on needs. You can think about additional coverages and ease of closing. Last but not least, Tom mentioned that it is crucial to do preliminary research in cooperation with an agency prior to market research. 

Emerce conversion & analytics 2024 hall

Reduce returns as part of your CRO program

Xeï Hulshoff - My Jewellery 

“Would these pants fit me best in size M or size L?” Does this sweater look best on me in pink or white? Tell you what, I'll order both and then I'll just return the lesser one.” A normal response from website visitors when purchasing retail products. For My Jewellery, all these returns are a high cost. After all, packages have to be packed and shipped even though you know in advance that at least 1 item will be returned. This causes more emissions by carriers, 2x costs for handling in the warehouse, etc. 

In this session, Xeï talks about how they are addressing both customer and business needs in reducing returns with My Jewellery's CRO program. 

As mentioned earlier, returns have a lot of impact on business, cost, not to mention the environment. So at My Jewellery, they started looking at customer wants and needs when it comes to returns. The usability study and a data analysis revealed that for new customers, postpayment is important. In addition, it revealed that some 15% of orders paid for in arrears are returned. And that if there is a cost associated with returns, it creates a negative sentiment because competitors offer free returns. 

My Jewellery then ran a test on product detail page (PDP) with the phrase “buy now pay later with Klarna.” This phrase had a positive effect on transactions even after correcting returns. However, it turned out that after evaluating on fulfillment margin, (packaging and delivery) which includes all costs, this phrase did not make money but actually cost money! 

Of course, My Jewellery was curious to know whether customers would really drop out if there was no Klarna option at all. They looked at this using an A/B test. What you might have expected: a heavy negative effect on transactions, even after processing returns. So it can be said that Klarna has a huge impact on transactions. 

Finally, Xeï shares some other insights: 

  • With free shipping, more items are returned 
  • Model and size information made no difference to transactions 

Reducing returns is an ongoing optimization and research theme at My Jewellery.

... Just a quick note: we send out a newsletter every three weeks that includes the latest blogs, team updates and, of course, news about the offerings in our academy. Click here to subscribe.


Newsletter sign up

Finding the heart in CRO: a guide to ethical practices

Mirjam de Klepper - UX researcher

In the world of CRO, it's easy to get caught up in data, numbers and the relentless pursuit of better statistics. Yet it's vital to remember that behind every conversion is a real human being: a unique individual with their own dreams, fears, problems, needs and desires.

Often, however, we are forced to think of people only in terms of points on a heatmap, click rates or as resources and capital.

In this session, Mirjam de Klepper discusses how we can incorporate a more humane approach to our daily work in CRO so that we can grow, not just as employees, but as people.

If we look at health care for a moment, people-centeredness is the absolute key. The ethnic basis here is simple: making sure people don't die. If we relate this to our field, we see that there has been a paradigm shift in recent years. From an old paradigm to a new human technology paradigm. Consider here that: CRO = a belief = power = responsibility. 

We are a lot more aware these days that our biases are creeping into the datasets used in technology. Think of AI in this context as well. Mirjam explains that it is important to be clear about what your goal is (intention) and what your impact is (result). Think about Airbnb, for example. Airbnb has a huge impact on the housing market, but this was not the original goal. 

Miriam admits that the current models still have little room for ethical issues. Therefore, she explains that steps need to be added to the current models: 

  • Evaluate after defining
  • Predicting after shaping an idea 
  • Monitoring after testing

Finally, Mirjam's tip is to implement failure mode and effects analysis during your work. This makes you think about how many different ways something can go wrong. Therefore, treat your departing customers as good, perhaps they will come back!

Sexual Wellness Insights: Engaging Product Presentations 

Camila Dutzig - EQOM Group

Camila Dutzig's presentation was the last presentation of the day. She conducted research on the potential of positively focused reviews in product presentations. And this was just as well: it led to a lot of growth and success!  

At EQOM Group, a consortium of major erotica companies, they came up with the idea to see what effect it had to capitalize on your customers' emotions to strengthen your product presentation strategies. When we look at selling technical products like a TV, specifications are very important. But in the world of erotica, exploration, pleasure, playfulness, inspiration and enjoyment are more important than specifications. In fact, customers spend much more time on such a site reading and checking reviews. That's why at EQOM Group they experimented with reviews on the product page. 

In the first test, they put the review directly below the product title. These were the reviews with the most thumbs up. They only showed the first few lines with a read more link. However, this had a big negative effect. So showing reviews (even if they are positive) does not work for every target audience.

Fortunately, failure is an excellent way of learning, so they continued with other tests. This is how they found out that the majority of users were looking at their site on mobile. Also, clicking on ‘read reviews’ increased significantly, which also increased time on the page. Super! 

Mirjam closed her presentation with a message we fully support: keep testing! Experiment with moving assessments around, show rules of assessments in other places on the PDP, and above all, keep testing. 

You might also find this interesting: