How to test growth hypotheses in SkyScanner

Transferring the experience of transition from traditional marketing to Lean & Agile marketing to SkyScanner from hopox.com service - project management based on Lean & Agile Marketing



As part of Skyscanner's transition from traditional marketing to Practical Marketing, we have transformed the development of marketing companies: from the very traditional big bang development process to an iterative process that is a great source of inspiration. Our final process followed the Agile Marketing principles manifest (http://agilemarketingmanifesto.org) under the influence of the Practical Startup (Lean Startup), flexible methodologies (Agile), the Growth Hacking mechanic, as well as some elements of traditional processes to do what is really right for our business.

')

The key idea is not to spend a week (or even months) on launching ideas and not testing ideas in artificial environments such as focus groups, then hope for success. We work on the principle of Lean Startup Create-Measure-Learn (Build-Measure-Learn): come up with an idea, create a hypothesis, choose a growth metric, develop a minimally viable company (not even a minimally viable product) and run it on a small segment of users to evaluate reaction, and repeat it until we find an idea ready to scale.

Process design

Our first iterations went pretty well under these principles and worked fine on small single channels, but when scaling up to more complex complex companies, there were too many variables, so it was difficult to correctly assess exactly what worked wrong. Was the design bad or were our channels not efficient enough? Or just the primary idea just sucks? Therefore, for integrated solutions, we combined the principles of a lean startup with traditional marketing issues:

1. Audit stage. Who might be interested in this topic?

2. The stage of testing the idea. Is the idea of ​​the chosen audience interesting?

3. The stage of verifying the decision. For the selected audience, is the realization of the idea interesting?

4. Stage channel verification. Does the solution work in all channels?





Old way of traditional marketing





Practical marketing, which is used in SkyScanner.



Each stage is a minimally viable idea, which is implemented at the time of its passage. In the early stages, you can test simple ideas on different groups of users, such as an idea with temperature-dependent content. The idea with the beach temperature indicator turned out to be viable to stimulate orders, but to check it out, we simply made a number of announcements and compared the temperature efficiency versus the lack of temperature.







At that moment, when we confirmed this idea, we moved on to scaling to other stories, such as food and travel guides. That this has already become a matter of allocating resources for scaling.

Test and test ideas

Each stage is a test, a comparison of two ideas, and there are a large number of options for further research. However, in the true Lean Startup there is no alternative, except to work in, albeit in a wild forest, but in front of real consumers, to see their reaction, measuring it with the growth metrics necessary to scale the idea. Experience shows that we spend 10% of the campaign budget on various tests. They should be available only to a few hundred people at a time, but you need to go through each feedback loop many times until you get it (the above Spanish example had 12 iterations over several days), but at the same time we should be ready to fail at any turn. We have many companies that have not been tested.

How did this work for us?

The speed and amount of work that has been done through this process has increased significantly. The last major campaign to launch new tools took about 6 months, until they entered the market, but recently a campaign of a similar size was launched in just a month, and some small scalable ideas in 2-3 days. But a real victory in the quality of the ideas that go to the market. What scales now has been tested in combat conditions on real metrics. We quickly made a mistake and found something that does not work in the early stages, thanks to the testing process, before we start to scale failure.



about the author

My name is Douglas Cook and I am the product owner at Skūscanner in Europe. Our team helps develop and implement processes, tools and technologies that help us move from traditional marketing to growth-oriented business. Having started my career as a traditional marketer, I fell in love and became part of the transformation that Practical Marketing provides, not only because of its influence on business, but because it constantly develops my personal skills and expertise, which I can translate into what makes Skyscanner so Great place to grow and study.



Original: https://medium.com/@Skyscanner/developing-a-lean-and-agile-marketing-process-d1822be47671