How does an analytics journey look like in a startup?

Start up was meant to be a temporary journey after which a shift to greener pastures was envisioned. Little did I know then that I would spend half a decade building some amazing things and experience growth from a remote warehouse to an office in central business district.

The aim of this article is on the challenges one might face in setting up analytics (start up particularly) and some possible solutions

On my first day, I had to learn a set of instructions from the person I was replacing as to how effectively I could "copy and paste" from different sources to form a report. A simple database existed which did not refresh most days of the month. Graphs or any forms of visualization was strictly forbidden, all reports were manual and hand made. An employee's productivity could be measured through average copy paste lead times.

With this situation the company is far from data driven, highly reactive,error prone, siloed, non standardized and many teams needed data to operate efficiently. The first thing that was needed was a platform to be put in place to achieve self service analytics. however being a startup which is cost conscious you cannot have all the weapons you need but however if we do decide to get some we end up with AK 47s in the era of rail guns.

The choice of weapons

On a positive note we did on board the AK 47, a cloud based business intelligence platform based on flash in the era of HTML5. which was significantly cheaper, provided dedicated cloud space, customer service to cater to tickets and training. It solved the job of self service analytics we were trying to aim, the data warehouse team worked round the clock to pump data into the cloud so that everybody could access it.

The most important part in the journey starts here, not all tools are meant for all companies. there are several parameters that needs to be considered other than cost. the ease of uploading data (in terms of volume) to the platform and updating it, refresh lead times, location (if cloud based) latency, price per user, yearly license fee, email scheduling, performing time series analysis, ability to input data into the tool, visualization quality, ability to export in different formats in volume, community, ease of syntax, modularity (ease of adding and removing a feature like extensions in tableau), ease of creating tabular reports, ability to standardize metrics and ability to connect to various data source(s). Did we look into all these metrics when we decided the weapon? of course not!

A for Adoption

no matter how much you believe in your product it will not sell itself as its not an iPhone.

This is the second biggest challenge startups face. the platform required training to operate, it had its own language (syntax) to follow and drawing data from cloud means a strong internet connection is needed which most of the countries we were operating in did not have and hence the tool was slow. After going through extensive training we had the job of extending this training to other stakeholders in the company, we also needed a platform manager to do the admin stuff of maintaining folders and schedules in the platform and above all this the first question that was posted to us was, "Can we download the data to Excel?"

A robust training program especially on the syntax, data sources,metrics and lastly how the platform can help end user in their day to day activity will go a long way. Adoption monitoring through KPIs, cutting off accesses to data warehouses and other querying platforms, the central BI team or the centre of excellence team does miniscule projects in various departments are some of the methods to ensure effective adoption.

Reporting monkeys

Despite all these shortcomings the tool was adopted and slowly spread in a unique way in which both excel and the platform coexisted as there was no other tool in place to quickly draw insights and help the business, implementing the cloud platform had standardized our reporting. slowly things began picking up, more and more people came onto the platform and started using it, drawing data and reports, then it results in what I call explosion of reports. we had several reporting monkeys churning reports and sending it to each other creating confusion in the business. this is another side of mass adoption we got to witness.

Establishing central of excellence team and associating departments with analytics will be key here, for example all the reports, dashboards and data product emanating from a customer service team should be labeled customer service analytics department and have their own workspace installed.

The flower blooms

Data became the buzzword of the century, price per user for business intelligence tools started falling drastically, we finally had a chance to move out of this legacy cloud tool and embrace latest platforms which were far more universal and used by top dogs in the industry and also much faster, lesser learning curve. for the first time in company's history we started embracing platform based solutions than reporting and excel based solutions, a consultative mindset crept in which data marts were established to answer business objectives and form the numbers need for strategy. the email schedules and reports started dropping drastically.

In all the above paragraphs an important part is missed i.e. people. Especially in this kind of environment people will start doing things on their own just because they can unlike in big companies in which its very much difficult to change things. Startups face a lot of attrition which is a double edge sword. any training program in place for adoption becomes futile and leads to business process being changed as the new person would conduct the business with a different perspective. The company has the best analytics infrastructure in place in the group today with most of the team having access to data to perform their duties, test various business questions posed and test strategies.