Most businesses now appreciate (if only conceptually) that data is their greatest asset. But despite a newfound attraction to becoming data-driven across functions (finance, operations, sales & marketing, merchandising/procurement) as well as the vast amount of budget being allocated to utilisation of that data, a large number of businesses remain unable to harness its true potential.
This is because many executive teams think that developing an analytics team is as simple as hiring a couple of analysts and awaiting insights into the meaning of life, the universe and everything. Unfortunately (although not surprisingly) building an analytics function is not quite that simple. This is mainly because there are a huge amount of tasks, processes and house cleaning that need to take place before data can be accessed and understood, let alone mined for answers to your business challenges. So when it comes to actioning analytics, the quintessential query you’ll need to grapple with is: to buy, or to build – that is the question!
In this article we outline the 4 different options you can choose from to create an effective analytics division within a business, explaining the team/capacity requirements for each.
Build your car – Drive it yourself – Navigate by map.
Starting with the full in-house solution, organisations need to recognise the importance of engineering resources to first extract, clean and join siloed data. Not to mention building data infrastructure to schedule and automate these tasks moving forward. This is the equivalent of your engine, without which all you have is a pretty dashboard that’s taking you nowhere – fast.
So once your data is in ship shape, your data analysts can build automated dashboards to track and assess your KPIs. At which point, (once you’ve built up enough historical data) your data scientists can build machine learning models to enable deeper insights and predictions to be harvested from the data.
Lastly, your internal business users need to be capable of understanding and actioning these insights!
Buy your car – Drive it yourself – Navigate by map.
If a business doesn’t want to hire engineering resources to build and maintain the data infrastructure engine, they may choose to outsource this part to a provider such as Conjura. This way they can keep their eyes on the road ahead by building strong internal data analyst and data scientist teams to operate their analytics.
As most of the failures in analytics functions come from inaccurate data or errors in code (read about these here), outsourcing this component of the analytics suite is exceptionally popular.
But don’t forget that once your analytics are up and running, your team will still need to know how to use that information to get you to where you want to go.
Buy your car – Hire a driver – Navigate by map.
Often smaller and sometimes medium sized businesses will choose to outsource the data infrastructure piece as well as the insights generation piece. This way they are able to really focus on setting up the commercial team to deliver value on top of a provider by navigating towards better business outcomes.
The final consideration with this option is to ensure that your team manually takes those insights, cohorts, data science model outputs and feeds them into other tools such as email, offline marketing and ad platforms – equipping them with the context to do things like send specific email messaging or discounts to the right customers, at the right times.
Buy your car – Hire a driver – Navigate by GPS.
In this fully outsourced model, data, insights and outcomes are fed back to various platforms through API integrations to perform automated actions. It is a useful model if you need to move at speed OR your internal team doesn’t have the capacity to perform manual actions off the back of insights from a dashboard.
This option gives you a working vehicle, with a dashboard you can read, being driven by a professional driver. All your commercial team needs to do is tell the driver where they want to go and then sit back and enjoy the ride.
It is worth noting that more advanced digitally native businesses tend to utilise automated and manual processes to generate value from data insights and would be technically option 1 with a mixture of option 2.