What Your Data isn't Doing Make sense: Five Things I Learned From Narrator Co-Founder Cedric Dussud -

2013. It's Cedric Dussud's first week in the role of the new engineering director at WeWork.

WeWork hasn't yet become the business that has transformed the business world. It's only managing a half-dozen office spaces in New York at the time.

However, it's increasing rapidly.

That first week, Cedric got called into Adam Neumann's CEO's office.

"So that's Adam, the department head I, too," Cedric told me. "And Adam basically says, 'Hey, I need for you to develop a billing system. And I need it in the next three months.'"

That's exactly what Cedric and his staff did. His main job over the next five years was the expansion of WeWork's billing systems in the course of growing the business from 8 locations up to 400.

In the process, Cedric learned a thing or two about handling information.

In fact, Cedric and Ahmed Elsamadisi an executive engineer in data at WeWork at the time, learned so much about what it took to scale the data operations that they devised a whole new system to make it simpler.

Together, they created Narrator Narrator, a platform for data that allows companies to run business intelligence without a large Data team.

Cedric was able to sit down with me recently to discuss the key lessons he learned about data at WeWork -the lessons he learned and how they helped to create the foundation for Narrator today.

There were a lot of ideas that stood out to me. Among them were:

• The reason why metrics don't always add together.
• What happens when there's no longer any one reliable source for truth.
• Why you should be skeptical regarding the privacy of your personal data.
• In the event that you need an expert in data science to assist you (and in the event that you don't).

You can watch the full interview below or just read some of the key takeaways I gleaned from our chat below.

5 Lessons Learned From Cedric's Experiment

I learned a lot watching Cedric's interview. Below are my five main takeaways from the interview:

1. Data Formulas Are Subtly Individually

Data science has a key difference from software engineering.

In the field of software engineering, when you work on an established system, it is possible to take what previous coders did before you.

However, with data, every when you design something completely new it isn't possible to duplicate the code and because of that, the definitions can have subtle variations across the reports.

Cedric employed the instance of a CEO who requested the quarterly report on sales, broken down into regions.

"You typically have to start over and create a new SQL to handle it," Cedric explained.

"You kind of hope that if someone else made the quarterly sales report in the past, that you will be able to find a way to use the report to split it by region, and maybe copy their logic. If you can't understand their reasoning, you'll have to create your own and your logic is going to be subtly different."

In the event that you're using these data for your company little differences will add up quickly.

"You will end up having hundreds of subtle differences within the organization," Cedric told us. "And it's very common for the CEO to come to the company and ask, "Why the numbers aren't matched? What's wrong?"

2. There Is No Single Source of the truth

The platforms that have built-in analytics will tell you a lot however they aren't able to tell you anything about data flowing through other systems in your company.

Cedric used the example of online ads.

A platform will give you reports on the number of ads run, the cost and so on. They may also attempt to estimate how many people have seen an advertisement before making a sale. But if you're running ads on multiple platforms, then there's no single platform that can provide all of the details to provide a complete answer, because they're not connected to one another.

"If you were to sum up all the orders that Facebook thought it was accountable for, as well as all of the orders Google believed it was accountable for, and so on, it would sum up to far above 100percent of your total orders" Cedric explained. "That's an instance where it's not possible to trust those numbers. The truth doesn't exist in any of them. The truth is that it exists in all of the other ones."

3. Do not be afraid to question your Metrics

Cedric suggests that you not believe in a company's analytics unless you know exactly from where these numbers are being generated and how they're determined.

In other words, you must dig enough to fully comprehend and are confident in the data that you're dealing with when you make business decisions.

"It is more harmful to overload data than it is to underdo it," he explained. "If you're thinking you're not doing it right and you're not doing it right, you'll engage your brain a lot and think twice about your actions; whereas if you've gotten some fancy system that answers every question for you and you'll think, Cool, I'm doing data."

4. Hire an Analyst to analyze your data when your Data is Spread across multiple Systems

There is no need for a massive analyst team in the beginning. Actually, going way too deeply into your data could be an unintentional distraction

"When you're a small company that's got your head occupied with how everything is going and you're able to comprehend the metrics you're using," he explained. And in that situation, you don't need an analyst -- in the moment, at least.

However, when your technology stack grows then the figures begin to be untrustworthy for founders and CEOs.

"When you're at a point where data starts to get spread across multiple places, that's when you lose the capacity to comprehend what's happening."

That's when it's time to engage an analyst, or to build out a more in-depth data strategy -- an expert who will help you fully trust the numbers again.

5. Concentrate your efforts on projects that will attract new customers.

Although Cedric created the billing system that runs a global business like WeWork however, he would notrecommend that SaaS founders should do the same.

It required months of continuous working for Cedric to put a preliminary version of WeWork's billing system into production, an effort that -- in his opinion -- does not merit the expense in small SaaS and software businesses.

"Work on your marketing. Work on understanding the customer as well as on new features- things that will bring the company new business," he explained. "Don't create your own system for billing."