In the first part of this blog “How Graze are experimenting their way to growth with Magento and AWS – Part (1/2)”, we made the case for Graze being the hallmark of an innovative company, by providing some examples of how they embrace the Lean Startup approach to quickly run experiments and validate they hypothesis, which can lead to new business models. This approach allows Graze to keep growing at staggering pace.
Graze’s experiments aim to answer the question “Should you build it?”, rather than “Can you build it?”. By rolling out Minimum Viable Products (MVP), Graze can learn what they need as fast as possible and at the lowest cost. Why? Because “The only way to win is to learn faster than anyone else”.
We know that we need to run experiments and learn faster than anyone else. But how?
The answer is in the scientific method, in “The Lean Startup” world known as the Build – Measure – Learn cycle.
- You start with hypotheses, a fancy word for “educated guess”
- Build a Minimum Viable Product, get it out into the real world
- Measure customers’ reactions and behaviors
- Learn from this, and use what you’ve learned to build a better product.
Let’s be clear. The goal of the process is not to get data, the goal is to get insight.
Let’s go back to the basic of the scientific method for a moment. In any experiment there are:
- Independent Variables, which you control
- Dependent Variables, which you measure
- And constants, that is everything else you don’t want to change
One of the challenges, as you keep experimenting to drive growth, is to make sure you don’t have undesired Independent Variables.
In e-commerce, for instance, you might want to find out whether Google Shopping Ads has a better ROI than Facebook Dynamic Product Ads. So you set-up two campaigns on both channels, making sure you have the right tags to measure traffic, attribution, conversions etc… To make sure you are comparing apple with apple, you introduce the least amount of variation between the two campaigns. However hard you try, there may be invisible, but important variations ready to bias your experiment. You would rightly assume that the the User Experience after clicking on the advertised products will be the same from there on, but often it isn’t. The users clicking on the Google Ads might experience a 2 seconds slower response time when landing on the product page, compared to the users clicking on the Facebook Ads. In this scenario, the Facebook Ads will show a better ROI, due to a unknown variable, which it was assumed to be a constant.
Uptime, Performance and Scalability are very important, but even more important is the consistency of performance, so that you can reliably validate your learning. This is why scientists use laboratory environments to have direct control over most, if not all, of the variables that could impact upon the outcome of the experiment.
You may be asking: yeah that’s nice stuff, but why on earth as a retailer should I be so concerned about this stuff? I don’t have time for these things, I need to run my business and worry about growth, and bottom line.
If it was difficult for your business to embrace mobile, let’s look at the future, the real unknown.
Today there are 5 billion devices, growing to 50 billion by 2020.
Do you have a sense for the magnitude of platform, infrastructure and business disruption that this will demand?
Your business needs to focus on experimenting and learning faster than anyone else, that’s the only way to win. This is the reason why Akoova continuously evolves, to help retailers to be ready for the future, whatever it might bring.