guest article by Bram Foubert, winner of the Best Service Article Award 2016

I just uninstalled a trial version of some statistical package from my laptop, and last week I cancelled a one-month free trial of Amazon’s Prime Video service. I admit I’m a free(-trial-)rider. I keep telling myself that this is what free trials are meant for, and my decision not to become a paying subscriber won’t hurt the service provider anyway. After all, the marginal reproduction cost of digital goods is virtually zero. On second thought, I should know better…  A company project and subsequent academic study taught me that such free-trial campaigns may not always be harmless to the company.

Some years ago, we were consulting for a European telco. They wanted to know whether their promotion strategy enabled them to speed up adoption of their recently launched interactive digital TV service. At that time, due to the presence of considerable technological switching costs, the company was vying with its main rivals to secure the lion’s share of the market, and its shareholders were pushing to gain a strong foothold as quickly as possible. However, there was some debate over the effectiveness of the company’s expensive free-trial campaign, which allowed customers to try out the TV service for three months, free of charge and without any further obligation. In fact, the company seemed to be split up into free-trial believers and nonbelievers.

So we built a relatively simple econometric model – in industry, deadlines are usually of a different order. Due to some technical issues, the final results were only ready the day before we had to present at the company’s headquarters. Much to our surprise, the results suggested that the company would have had more adopters if it had not run the free-trial campaign. I remember us slightly panicking over this unexpected insight: saying that a promotional campaign hasn’t been quite effective is one thing, claiming that it has been bluntly counterproductive is another. Moreover, given the simplicity of our model, we could only conjecture what was going on. Still, we stood behind and successfully defended our findings. And the company ended up discontinuing its unconditional free-trial offer.

I like company projects but they often leave me wondering about two related things: why do we find what we find and to what extent can we generalize the findings (or, said differently, what are the results’ boundary conditions)? So we set up an academic sequel to unravel the mechanisms underlying the free trial’s negative impact and identify conditions under which the promotion would have paid off after all. This time, we developed a substantially more complex econometric model, with all the bells and whistles (such as learning effects and word of mouth). It enabled us to shed light on a crucial trade-off. On the one hand, free trials can attract new paying subscribers, by allowing consumers to learn about the service through own free-trial usage or word of mouth. On the other hand, due to a disappointing trial experience, consumers may decide not to adopt the system and, importantly, may be lost for good. Of course, this dilemma is particularly worrisome in early stages of the service’s life cycle (like in the studied case), when service quality may still be suboptimal. However, counterfactual simulation also taught us that managers can avoid such negative results through appropriate promotion timing. The trial promotion should take place after the service has been sufficiently tried and tested in the field to ensure a better trial experience and decrease the risk of prematurely killing customers.

Frankly, this more nuanced perspective on the effectiveness of free-trial promotions won’t keep me from enthusiastically continuing subscribing to free trials. In fact, I still need to figure out whether it’s possible from Belgium to subscribe to this one-week trial of the beta version of Hulu with live TV. If it works decently, I might even end up paying this time…

Bram Foubert

Associate Professor |  Department of Marketing and Supply Chain Management 

Maastricht University School of Business and Economics