No Bullshit BI
Trust - I feel there is a lack of it between two sets of people inside companies: Business people, salespeople, marketing, management, and the data people, data engineers, scientists, and analysts.
I often see that data people think businesses aren’t paying enough attention to data, while business people tend to believe data people aren’t focused enough on the business side to be helpful with their data.
I think we need to restore this trust, the trust that both sides care for the other enough to move things forward together.
I don’t think big talk or fluffy videos will do the trick. So I’ll try this: a simple set of ten rules I follow and suggest others follow when it comes to business intelligence - The process of guiding business actions with the help of data.
Take this as a first humble step towards a better relationship.
I believe two fundamental pillars underpin a successful approach to business intelligence, and those are:
Business should be first. Business should guide business intelligence in all its parts. Never the other way around.
Data is complex, and will stay that way. Not even great business people or strategy can make data simple. So we need data people to help guide business efforts.
A no-bullshit approach to BI built on these two pillars means choosing…
Actions over Data
Whatever we do to help people get data, we must always choose the path that gets them to take action quickly. If you’re a salesperson comfortable with Excel, then please use it! Data people should help you by providing workshops and CSV data exports so you can take action quickly!
If you’re a data person, learn about the tools business people use, understand them, and dive deep into their decision workflows to see where you can actually help and facilitate.
Complexity over Simplicity
Data isn’t simple; the world isn’t flat. Yes, you can make things simple(r) to understand by breaking them down, but you’re not going to remove the complexity! A Dashboard should indeed be complex, not simple.
If you don’t have the balls to dig into the data, you’re not gonna make good decisions, period. Don’t try to dump down data, instead business people need to dig deep into data. And data people need to teach others how to dig, and provide good shovels.
Enablers over Technology
Data people should be enablers first, and technologists second. They should enable others to use data, not scare them with technology. Business people need to stop looking into fancy marketing automation tools, and start understanding their own decision making. They need to help data people understand how they make decisions, so they can be enabled.
Fewer Decisions over More Decisions
Both business and data people need to come together and focus all the data efforts on a set of key decisions. The ones that truly move the needle for the company, and only the business side can tell what those truly are. But only the data side can provide sufficient data to actually move it.
Investment over Manpower
Data is a business decision just like any other. If you need to build a tracking tool to collect data, that’s an investment decision. If you need to buy a data set, it’s another. All data is an investment decision, and no decision is made by throwing lots of data people at it. It’s the data that counts, so be prepared to invest in it. Business people need to have the guts to spend serious money on larger projects or purchases, and data people need to expand their horizons to include the “non-DIY” solutions into their option space.
Usability over Feasibility
We should look for usable solutions first and feasibility second. That doesn’t mean ignoring feasibility; all solutions have to be feasible. It means we prioritize usability first and then search the usable set of options for the feasible ones.
We must prevent them from ending up with a feasible but less usable option at all costs. A lot of companies try to force every employee to learn SQL to work with data. That’s a mistake, a mistake that excludes people by not being usable first.
Decentralization over Centralization
When we use data, we want to use it decentrally. Unless you’re a startup, your workflows will be decentralized, inside teams, inside departments. Decision-making is decentralized, and so should your data usage.
That doesn’t mean you have to adopt the data mesh approach, not at all. But it does mean you should think about reverse ETL. This means that you should think about having analysts inside departments; whatever your company needs to use, data should be readily available wherever your decision workflows happen.
Near time over No Time
When we can get data faster, we should. Data people should try to get it to business people fast, and business people should try to look at it quickly.
Speed over Process
When we need data urgently, we should have it. Speed is of uttermost importance in decision-making. We should never have a process in the way of getting data.
Added Business Context over Big Data
We often need to choose between adding more data and adding business context to the data. Both data people and business people should always favor business context over sheer amount; it will always pay off in better decisions down the line.
The Company Over Business and Data
I think both data people and business people need to realize that neither of them comes first. The company comes first. It’s not a matter of figuring out who should be prioritized inside the company. Instead, we should all aim to figure out what the bigger goals of the company are, and see how each business function and data function can support those.
In that way, we can align together and build the trust we deserve.