How startups can use data effectively

startups data dashboard

“It is a capital mistake to theorize before one has data”. Although not my favorite one, this quote, assigned to the fantastic character Sherlock Holmes, summarizes why the matter here discussed is so important. Data, if orderly collected, correctly treated and well analyzed, leads to final, definitive truths, capable of guiding businesses to the fulfillment of their potential or, at least, to the avoidance of their disaster and ultimate bitter end. On the other hand, without it, even the most experienced leaders are blind and at a real disadvantage when making decisions – some that can be fatal when wrong, especially considering the super-competitive era we live in today.

In ambitious companies with fast growth, handling data correctly is even harder. The number of gigabytes that quickly accumulate in the warehouses can be overwhelming, the infrastructure to efficiently query them repeatedly insufficient, and the expertise to draw useful insights from the different sources lacking. Thus, from the very beginning of a company, it is important to think about the different ways in which data can be used to make the business more intelligent, in every level of management, and how to get there. Here are some tips on how to do it:

  1. Setup a solid data infrastructure for your company:

    Although changes and migrations are often necessary, having an intentional stack of tools to construct reliable data pipelines, almost always ideally self-sufficient, dynamic (capable of rerunning and refreshing periodically) and scalable is really important. Thinking about permission layers as part of the system is essential as well. It should be able to easily determine what each individual can access and do, which can be based on the concept of data roles inside the company.

  2. Have clear guidelines around data usage.

    Given the necessities of privacy and security, besides having a good infrastructure, it is important to correctly manipulate the data available, giving the right level of access to the different contributors in the company, explaining what it is and why. Besides that, it is important to clearly state what can and cannot be done with company data, shared or not, and in which ways with which people. This is particularly important for individuals dealing with business partners, advisors, outside services providers, and other third parties. Not only by having written docs but also ongoing training and talks about the topic is the right way to go about this.

  3. Be as clear as possible on data peculiarities and offer orientation for data analyses.

    Once people get what they need for their tasks and know the limits of what they can do with it, it is then necessary to extract the right meaning from the data available and understand the implication of what they are looking at. Nuances about similar but still different sources of information need to be clear to avoid confusions and mistakes. The purpose of various filters, which can be often forgotten / neglected, thus leading to mistakes, need to be emphasized. Finally, although the idea is to make individuals as independent as possible, extra ongoing orientation when needed has to be made available by data teams.

  4. The most common analyses should be saved in a systematic way for reproduction, and data visualization tools explored for extracting insights.

    For the most frequent analyses made by individuals, repeated queries and data pulls should be saved and made available for those who use it, so no one has to repeat work recreating them, and also can be sure that such way is the right one to get certain key information. Also, data visualization tools, often underestimated, should be heavily explored – graphs and charts when well thought out, are usually more meaningful than raw numbers, especially when they are simple and straightforward.

  5. Team KPIs should be shared with all team members and company targets with the whole company.

    Especially in the form of dynamic and simple dashboards, team performance goals, when shared, lead to a strong sense of ownership and orientation towards a common purpose – which often is associated with better results. The same is true for company targets shared with every single employee. It is important to have focus here though – too much data can create confusion and lack of focus. Five to ten measures/numbers – summing up company and team ones – is probably the ideal.

With these tips, we at Brex believe you can also have world-class data in your company. They reflect the philosophy that we also used to build our own systems, processes and organization around this precious resource. Best of luck on the journey to get there.

Keep growing!

Related Articles

blog footer
Introducing the Brex Quarterly Review
blog footer
Brex talks: when and why to launch your second product
blog footer
How Not To Be A One Hit Wonder: On Building Your Second Successful Product
blog footer
How we chose our third vertical: Brex for Life Sciences