In God we trust. All others must bring data – Building analytics in to the system – The Why and How
“In God we trust; all others must bring data.”
This quote is widely attributed to many people. However, in indian folklore, it is NRN who said this.
As Sherlock Holmes would say “Data.Data.Data.It is a capital mistake to theorize before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts.”
For a Startup,for that matter, any product/service, data is everything. It will tell you everything about the past, present and future in a brutally honest way. But the real trick is in building analytics into the system to obtain data and consuming it.
In a startup where you operate in a hypothesis driven model, you have made lot of assumptions about your solution, customers,pricing and a whole gamut of things. Looking back, you might have experienced one such following situation.
a. Sometimes you build a feature with lot of certainty and it will not work out the way you had fancied. In few cases, you will not be even aware of its failure.
b. You would think that you have done everything right but still things are not going great guns.
c. Sometimes you would think all is well and you are on the right track, but suddenly you hit a dead end.
This is not exhaustive list, but you get the drift. All these could have been avoided or at the very least, dealt in a right way. How?
1. Build Analytics into the product: These days, every startup uses Google Analytics, Kissmetrics etc. These state-of-the art tools offer range of services and options. Metrics such as number of unique visitors – daily and monthly, drop offs from each page,search traffic, user navigation, heatmaps etc are part of their offerings. They are very important in the scheme of larger things. But it will not suffice. You have to build analytics in to the system.
2. Before embarking on the quest of development: Before starting building a product or next version of a product, you should start by asking lot of questions. They might be about some assumptions that you have made. They might be on your potential revenue channel. But the point is, you have many questions/doubts/assumptions and you have to list them down,meticulously.
3. Find the right tools: Now, segregate all those questions/assumptions/doubts that can be found out by existing analytics tools(your Google Analytics and Kissmetrics). Rest, you will have to find out by building them into your product.
4. Pick your battle – weed out the data: Once you have setup right tools and built analytics into the product, the next thing to do is to collect the right data. Sure – the more, the merrier. But it is a timesink.You will have to weed out all that is not important/makes no sense/false data. Otherwise, your decisions based on them will be skewed.
5. Post release of your product: Massive data will be in your kitty.You have to look at them and find your answers. Educate the team on the data and decisions taken based on data.
6. What next? Now that you have answers to your questions,you can plan your next iteration with more clarity, certainty and focus.
Though every product/service will have its own unique questions and metrics, there are some fundamental questions that should be answered by analytics in every iteration of your product.
a) Who are the customers? Where does my traffic come from? What is the customer acquisition cost?
b) What are the customer segment(s) that is to be targeted? or What are the customer segments that are already being targeted as part of this release version?
c) How to better serve my existing customer? Is their real problem being solved by this iteration? Is the customer really getting the intended value?
d) What are the problems that are faced by a new customer segment? How to solve it? What is the intended value proposition? Is the customer aware of it?
f) How to make an user to repeatedly use the product?
g) Who are the real users? Actual users are set of users – who are very real, who are using your product/service, just because you are solving their real world problems and who are ready to pay for it.
i) What is the goal of this iteration of product? How to measure the success of this iteration? E.g. Your goal might be to acquire customers in first iteration and acquire/activate customers in subsequent iteration.
h) How are users really using the product? This is very important. We have assumptions on how people will be using our product/service. Suddenly, we might stumble on a completely new use case which was not thought of in the first place.