Over the past 3 years I've worked with more than 100 companies like Subway, Ab Inbev, Telefonica, Rappi, KIA, throughout a wide range of industries like B2C, B2B, E-commerce, Packaged Goods, Airlines and Retail. I’ve seen first hand their data approach and the way teams are organized around it. In this blog, I will be sharing some thoughts and learnings on how small and big companies organize their data teams to excel.


In general, companies struggle to determine how to organize data teams. And how they should work with other areas within the company. The following questions are common in this process:


Should BI be part of Marketing?

Should BI be an independent team?

Should I have a BI team? How big?

Should I have an external team?

Should we build the tools in house?

Do we have the ability to build them?

How can we explore Artificial Intelligence?

How much is this going to cost?


With this in mind, here are the 4 key pillars I think are imperative when structuring a data team:


1. Structure:


I believe BI shouldn’t be part of a specific business unit nor should it be the team “in charge of data”. I believe BI should be a support unit to serve the organization horizontally. Based on this approach, BI can be in charge of:


a) Providing data-focused training to all decision makers.


b) Supporting technical issues when these arise. For example: when running an SQL query, etc.


c) Making sure all teams are using software correctly.


d) Ensuring each area is using data to make decisions. 

*Note I am not proposing BI teams to be in charge of running models and administering software or data... it should be each department’s responsibility to do so.

‍

2. Technology:


Decisions on which technology should be adopted across the organization should be made by the CEO, CIO/CTO, BI and CMO. Nowadays, these are decisions made by each unit, and I differ on this approach. Technology is a strategic move and now more than ever, with collaboration between teams as one of the most important factors. According to a recent survey made by Deloitte (Source: Deloitte´s 2019 “Becoming an Insights Driven Organization” Survey) companies that use the same system across the organization exceed their goals by 80% vs 60% from the ones that don’t. That’s a whopping 20% increase which can dramatically change how companies perform.


3. Internal capabilities vs external:


When we understand the complexities around data, which include infrastructure de-duplication, analysis, predictions, execution and people, I recommend staying away from building software from scratch. Also, I believe it is a big mistake when companies close their doors to new technologies. In some cases, companies want to shift into a data-driven culture from being a product-driven company, like it is the case of Nike. Their approach is geared towards creating internal teams that build around data in-house, and even though this sounds like a good solution, it could be considered somewhat of a trap. The reason for my reasoning is because these types of companies already use either Azure, Google or Amazon to serve their infrastructure. Leaving them tied up to these single sided softwares. Technology changes every day and closing doors on new technologies or new ways of doing things can be a huge mistake. I have had cases where directives within companies tell me “well, this looks awesome but we’ve been working on our own system for a couple of years now and at this point I can’t change it”. This happens when they have made a bet on building software in-house, so their internal teams can’t say that they found something better and cheaper in the market - for political reasons. Technology is precisely about being able to change what you have, it is about iterating, it is about trying what’s new and adapting. Otherwise, what took 2 years to build, will end up being useless.


4. Costs:


As expected, building internal teams and technologies end up incurring extremely high and recurring costs, and a feeling of dissatisfaction with the system's performance. Integrating data with companies like Salesforce and IBM result in endless implementation and consulting efforts which are typically led by BI teams. Unfortunately, these projects result being frustrating since only BI teams have control over its data. Until recently, it was advised to think small and solve a specific problem with one technology, which learning from my experience, this causes more problems than solutions. With companies like Intercom.com and Notion.so, things slowly started to change thanks to their all-in-one product initiative. In Intercom's case, for example, sales and customer teams can be unified. Based on this, it is recommended to think about strategy and how must the company work and behave.


My advice is to look at the big picture and avoid looking for a product that solves only one specific issue. Products that offer a full service are solving an ecosystem of issues within organizations, and that is because the era of hyper specialization and segmentation is over. We are living in a complex world and systems must be unified to ensure optimal efficiency. In the end, looking at the big picture and taking the time to do research will prevent endless costs for the organization.


To sum it all up, build data teams that can use your data but one that doesn’t own it, rather make it a support team for other areas within the organization. Make data teams teach others how to use technology and data insights and how it can be used to exceed their goals. Don't try to build technology in-house, be open to new approaches and new systems that could make your company perform better. Reach out, seek help, keep your ears open. Finally, do not buy technology just thinking about independent departments like BI or Marketing or Sales. Think about technology as a key element in your strategy.


For more information on datagran visit www.datagran.io‍

What's Datagran? Datagran is a system of record that allows companies to integrate data and analyze, predict and take action in a single pipeline to solve specific business problems like churn, increased inventory value, customer acquisition cost, and more.