5 Mistakes Startups Make With Data And Analytics

Startup founders are constantly being told that they have to build their data and analytics. They hear it from board members and other entrepreneurs. And they see other companies using the information to make better decisions and build more personalized products.
‍

Many try to take the advice, but they don’t really know, how? And time after time, I see startups making choices that will make it far more difficult for them to use data to succeed. Here are the top mistakes that young companies make:

‍

1) Hiring a data analyst upfront

Bringing in an analyst can seem like the most effective way for a small company to start using its data. They can grab reports from your various systems and perhaps build a simple dashboard. This way, you figure, you can put off investing in a full data infrastructure until the value is clear.
‍

This may sound smart in theory, but it doesn’t work in practice. If all you have is individual systems without a core data infrastructure, everything the analyst does will require a lot of time-consuming manual effort. When you do decide to do it right, you’ll have to throw out all this work and rebuild everything from scratch.
‍

What’s more, using an analyst actually makes it harder to build a data-driven culture. It’s much smarter to deliver self-service analytics so everyone in the company can interact with data and answer questions directly.

2) Putting off building data infrastructure

Sure, it may seem daunting for a young company to build the infrastructure to collect, store, and analyze data from all their applications. Wouldn’t it be easier to put that off for a while and simply rely on the analytics built into each system they use for marketing, product, finance, etc. Maybe they put together a few spreadsheets to pull some figures together.

This is shortsighted, and it will wind up being very expensive. I look at core infrastructure as the foundation of a building. If you don’t get it right, your skyscraper will come crashing down. The core layers of data infrastructure–tools to ingest, store and analyze your information–are the foundations of any business in the digital age.

3) Underestimating the challenges of building from scratch

Maybe you decide to build a data platform, but you figure you can hire a smart but inexperienced engineer who will be able to figure out what to do. You can’t. There is a lot of skill and experience that goes into building and maintaining a data infrastructure, not to mention making sure it has the security, privacy protections, and legal compliance you need. It’s rare to find a junior engineer with knowledge of all this. After all, a large technology company may have 50 or more people devoted just to their data platform.

4) Outsourcing infrastructure and operations to consultants

A technology development shop or consulting firm certainly can set up the data warehouse, analytics, and other functions you’ll need in your data infrastructure. This will cost a lot of money since consultants bill you for starting every project from scratch. In the longer term, there’s an even larger cost: Your team won’t have the expertise to maintain and expand your data infrastructure.

5) Starting with machine learning or other advanced applications

There are a lot of powerful applications that can be created using artificial intelligence techniques that will likely propel your business forward. But the dirty secret of most AI projects is that most of the resources are spent cleaning data that is fed into the machine learning systems. Investing first in the infrastructure needed to collect consistent, accurate data will make your advanced applications faster and more powerful when you get to them.

Reading this list of what not to do may make you ask if there is any way a startup can get data right at a price it can afford?

Yes, there is. It starts with the realization that you can’t afford not to have a solid data infrastructure from the start. It would be just as foolhardy to construct a skyscraper without a solid foundation. That means getting the right stack of software and engineers who are skilled in using them.

We started 5X to make it easy for companies of every size to enjoy the benefits of the modern data stack without the complexity.  We’ve developed a managed service that includes everything a startup needs—first-class tools like Snowflake & Preset, on-demand access to trained engineers, and expert guidance from technical project managers to guide you on best practices. It’s everything you need to grow your company and avoid time-wasting mistakes.

‍