arrow-left arrow-right brightness-2 chevron-left chevron-right circle-half-full dots-horizontal facebook-box facebook loader magnify menu-down rss-box star twitter-box twitter white-balance-sunny window-close
Data Analytics is a Journey
3 min read

Data Analytics is a Journey

Data Analytics is a Journey

It is 2020 and the data analytics has gained so much attention even outside of the tech community. "Data is gold", they say - no one wants to be left behind. However, getting the right strategy is neither a straightforward nor static process.

It's not a Fiat!

Three years ago, one of my ex-colleagues said in his presentation:

What we have now is a Fiat 500. It starts, it runs, but it takes forever to reach the destination. We need to turn it into a shiny car.

It was a good analogy though quite an optimistic one (we had a tricycle at best, in my opinion). To create a good strategy, the first step is to understand the company itself.

Photo by Giorgio Trovato on Unsplash

Assessing the current analytics capabilities and drawing up the data landscape can be good starting points for creating a time-bound strategy with measurable outcomes. Getting the sequence correctly is also critical for any strategy. For example, there is not much value for a good data platform if the business does not welcome the culture of making data-driven decisions or the organisation is still illiterate about data analytics.

Point 1: Understand the landscape, accept the reality, and nurture a good data culture.

Towards the Horizon

It is a common sad story that random analytics capabilities are grown without clear objectives. For several, the thought of missing the trend has brought them into an unknown territory that costs millions. Formulating a clear vision should be at the top of their to-do list instead.

Rapid technology dynamics and new techniques for data analysis have made a long-term vision difficult. What would be more effective is to create a vision that envisages as far as the horizon. Defining high-level goals in the early stages to support such a vision also helps minimise the deviation caused by conflicting priorities.

Point 2: Set a clear vision and goals.

Data Teams Assemble

Explaining how to build an effective data team is like begging for troubles. Managers have their leadership and management style. Having said that, establishing a team to tackle data challenges requires some common components and to some extent, common sense.

The Enabler: Many companies try to create a complete centralised team with a wide range of capabilities from data ingestion to data visualisation. This type of formation is most common when an IT department tries to own the data space. IMHO, the approach often fails due to the lack of domain knowledge, demonstrable value and agility. It also creates tension between departments as there are no clear roles and responsibilities defined.

Whether the centralised team is brewed from an IT department or branched out from the CxO office, it should act as an enabler rather than focusing on the last-mile delivery. The team should be responsible for automating high-quality data pipelines, building and operating the data platform, and adding features that help the business gain additional insights.

The Driver: Successful analytics is driven by clear objectives defined within the business context. The "Big Questions" should come from the leadership team in the respective areas and must align with the business strategy. As domain experts, the business departments are responsible for driving the analytics agenda using the data available to them.

Point 3: Define clear roles, responsibilities and ways of working.

Are we there yet?

I wish! There is no silver bullet in the data analytics space as there are too many factors to consider. However, below are a few points that might help you avoid some basic mistakes when embarking into this data analytics journey.

  • Hire more techies than managers (1:1 or 1:2 is a ridiculous ratio).
  • Promote 'A', grow 'B', and stop hiring 'C' players.
  • Collaborate across departments. Better insights are gained that way.
  • Overfitting data models to reinforce your belief is worse than doing nothing.
  • Focus on your priority. The balance between value and capacity is important.
  • Take technical debt seriously. It tends to be exponential.
  • Compromise data quality in pursuit of speed is not going to get you anywhere.
  • Ignoring security is a recipe for disaster.
  • Context is everything. Review your data strategy constantly.

These may look like common sense but it happens in real life. Remember that a data journey is a marathon. Cruise your team at an optimal speed is going to help you sustain that constant delivery long-term.

Note: My friend Marco Casassa-Mont pointed out that the image above is a Fiat 600 and not a Fiat 500. Many apologies to all the Fiat lovers out there.