Distinguishing the signal from the noise requires both scientific knowledge and self-knowledge: the serenity to accept the things we cannot predict, the courage to predict the things we can, and the wisdom to know the difference —- Nate Silver
Any startup or any corporation for that matter has only two jobs :
- Knowing what to do
- Figuring out how to do it
A startup typically runs lean, at-least the successful ones while a corporation has the whole middle layer bridging the two aspects of what and how.
Given the understanding, analytics has a role to play in both what and how part of running a startup as it evolves. The kind of analytics needed will evolve with the business.
A startup journey from a strategic perspective will have 3 stages :
- Analytics at this stage will largely be part of the product team and more about what.
- The kind of work required will mostly be around defining metrics and ensuring that the direction for the startup is right.
- All this can be achieved using a basic set of tools involving google analytics, SQL, Excel and customer surveys.
- Keep it simple and make sure all decisions are backed up by data where ever possible: design, product decisions, feature priorities etc.
- At this stage, the start has largely figured out what is working and what’s not. Also, with a large incoming business, substantial data would start flowing.
- This will be about building for how part. Data engineering will be key at this point in time, Kafka, Luigi, MongoDB etc. The data pipelining needs to be more robust compared to earlier.
- The engineering team will have their task cut out in trying to store and make sure the data is in a retrievable format. Analytics will be mostly around making sure key reports are generated fast, consumable format.
- Analytics at this stage is still restricted to the strategy part but more from operational effectiveness and streamlining of business.
- The business has grown past startup stage with at least one stable product or a set of paying enterprise customers.
- The analytics now will now be about improving the top-line. This is where advanced analytics or data science will be taking over, product features like recommendation engines, AI assistants etc will be taking over.
- At this stage, analytics will also have evolved to have a separate division different from product and engineering.
- Amount of mathematics involved will increase, the goal now is to primarily differentiate and really innovate.
Irrespective of the kind of business B2C or B2B, analytics will and should evolve with the business. The backbone for strong analytical practice is solid product thinking and engineering practices. Without those in place, any firm will fail to leverage data science well.
As reiterated above, it should be clear where and how analytics is being used and can be used. In the initial stages, what should be the focus? As things get streamlined and the direction is set, analytics can be useful in the how part as well.