Lean Startup Analytics and Data Warehousing

Tom Breur

9 July 2017

Lean Startups need to move quick. In today’s data-driven economy, Lean Startups also need data, preferably lots and lots of them. Oh, and they need ‘m fast! The Data Science profession is making inroads in ever more areas of business and life. We are all familiar with GIGO, and awareness is growing that data scientists’ effectiveness hinges mostly on availability of clean and neatly organized data. The more the better.

Very few organizations can manage and organize all of their data in data warehouses, at least in an environment worth of Bill Inmon’s definition for it: “A Subject oriented, Integrated, Time-variant, Non-volatile collection of data in support of management’s decision-making process.” So more often than not data scientists will combine traditional with less refined sources of (raw) data. In fact, in many cases the decision to invest in further refinement of data sources will only come after commercially viable usage of data sources has been shown.

As I wrote in another blogpost, the “Cost of Delay” is a vicious feedback loop, and negating the consequence of delays may well justify incurring a fair amount of technical debt. Especially when there is an urgency to get to market fast (which holds for every start-up, of course). But when you know you are competing in a data intensive market, you may start small but need to think big (ahead): structured and centralized data storage will become a success factor. Building two or more initial iteration can be a good idea, just don’t grow too attached to your early brainchild.

Nowadays, many start-ups are launching their products even before there is any product. Talking about a minimally viable product (MVP) – no product at all! A product concept gets offered through a website, and the behavior that prospects display towards these prototypes is extremely valuable data. It will tell you very quickly what people like what they are interested in, and if you design it well, what they are willing to spend their money on.

When Tesla first announced their Model 3 in March 2016, this immediately drew enormous interest. Many prospects were in fact interested in paying down $1000, just so they could get early(ier) access to production models – and nobody had seen one yet! Marketers Walhalla: the dialogue with prospects must have been an extremely valuable source of information on desirable features, and willingness to pay. It is no coincidence that nowadays many products get launched (very) early like this: it provides extremely valid input from the customers that are most important to you – the ones willing to buy your product.

These marketing tactics enable you to “show” an MVP even before the first one has been built. Of course the value of these tactics hinges on availability and usefulness of prospect that are browsing your site. That is where your “back end” (data gathering mechanisms) will need to provide a MVP data warehouse. Agile Analytics, genuinely Agile data warehousing. Start small, think big!


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s