The Five BIGGEST Data Mistakes Business Leaders are Still Making Today

Oct 08, 2023

Having spent over 35 years in data analytics, mostly in the business setting, I have seen the good, bad, and ugly of how companies use data. Unfortunately, examples of ugly have outnumbered examples of good by a large margin. And today, most companies still get it wrong.

Below, I highlight five big mistakes leaders continue to make. Most revolve around the adherence to a compartmentalized view of their organization. They assume that operations are separate from human resources, benefits are separate from performance, and finances are separate from recruitment. They are not; each and every aspect of a company represents a puzzle piece within a bigger picture.

#1 – Resisting Integration.

Every company above 500 employees (and perhaps smaller) would benefit from internal analytic expertise about their own business. Period. I guarantee that the first set of insights a CEO sees from a fully-integrated view will alter their understanding of their own organization, permanently. Some of their realizations will be simple (true story: a CEO turns to the head of HR – “is that really how many people we’ve employed in New Jersey in the past year?”), others more complex (seeing that locations with higher-paid, more expensive workers more than made up for their higher expenses --- e.g. absence, benefits --- with greater net revenue).

I believe running a company without an integrated data platform is corporate malpractice. It is the same as piloting a plane on a cloudy night without instruments. Perhaps you know the general direction you’re headed and where you hope to land, but you remain under-informed about the current status of the flight or what problems you might have.

In today’s cloud-based data environments data transfers, integration and storage are straightforward and feasible. Security and privacy can be safeguarded with appropriate protections. And vendors have become efficient at importing regular data feeds. So, when you hear a leader say they don’t have the budget, that data integration is too difficult, or privacy concerns make it impossible, don’t believe it.  Perhaps those leaders don’t really want to know, or they worry they might look bad. Trust me, no company can afford to fly blind.


#2.  Relying on industry trends and benchmarks. 

I’ve known many brilliant business leaders who are not experienced with data analytics. Some surround themselves with people who can dig into their company’s own unique data and translate the discoveries.  Others avoid investing in their own analytic capability and rely on outside consultants and industry experts – presuming that those answers apply everywhere. 

Those who rely on industry trends often prefer the comfort of authority (“Mr. Expert says 50% of the fortune 500 do it this way”) to the uncertainty of what they might uncover when they dig in. It requires a certain level of courage to actively understand problems that may not have an easy answer. While some leaders say they want to be data-driven, not all of them mean it.

Having data means you’ll face the facts. Perhaps certain locations have a much higher rate of accidents, or that one department has low engagement, or that new hires are quitting more than you expected. If you’re not ready to own what’s happening, it’s easier to pretend that experts know more about what your company needs than you do.

The same goes for benchmarks.  Benchmarking is overrated. It helps – . Just because a data integration service has 100 million lives in it doesn’t mean it is the best platform for understanding your company. Once you understand how you compare to other companies, the next step is always to incrementally improve within your own. 

Decision-makers quickly learn that having an intimate understanding of the drivers of their own experience is far more important than knowing more about other companies. Plus, in my experience, the vendors with the most comparison data are also the least flexible. Serving as a benchmarking service means focusing on a few standard data sources and getting large volumes. This means they are less focused on the important idiosyncrasies of each unique company. Show me a leader who chose a data platform primarily because it could do benchmarking and I will show you a leader who is pulling their hair out 18 months later.


#3.  Allowing artificial separation -- people, business and benefits are all part of the same picture.

A company needs the full picture – its people, its operations, its spending, and its outcomes – not isolated, separate pieces.  Continuing our plane analogy, it’s difficult to make real-time flight decisions if you have to make calls to the head of altitude, the manager of fuel, and the VP of speed to get their independent input. 

It is only through integration that companies learn just how interconnected everything is. Turnover is function of work schedule, culture, engagement, compensation, and performance.  Absence and injury are a function of health, burnout, training, experience and policy.  Net revenue reflects how people are performing as well as how they use benefits. It’s one big puzzle, not a collection of little ones.

If I had a dollar for every time I heard that department X (safety, benefits, finance, etc.) won’t share “their” data, I would be able to buy a Tesla. It is natural to protect one’s turf. But departments don’t own data, the company does. Integration advances collective understanding for the good of the full organization. A department head reluctant to cooperate needs to be nudged out of his or her safe space.

This also means that the data integration platform MUST accommodate and encourage importing of all important data sources.  A vendor that does People Analytics but ignores health or does health care analytics but has no experience with business data will not produce an integrated view. Beware of vendors that have a few standard data feeds and then charges a hefty fee for each additional feed. If they fully integrate at the company level, there will be MANY examples of how they routinely incorporate non-standard data feeds, with no penalty for including data sources unique to your organization. 


#4.  Asking the wrong question (or assuming the usual answers)

It is natural to fall back on one’s own (narrow) experience to explain a problem.  For example, people trained in health will assume that high health care and disability costs are the direct result of greater levels of illness and injury. It’s not until they see significant differences in cost related to policy, economics, location, engagement, training, or experience that they widen their view of what constitutes a “cause.”

As described above, companies are complex systems that must be understood as such. I have seen hundreds of counterintuitive connections (sick leave related to compensation structure). Some had very straightforward solutions (policy change), others required more tailored adjustments. But all of them saved us from assuming (incorrectly) how to fix a problem or continuing misguided efforts that did not address the root cause.

Once data are integrated, the investigative team should include many stakeholders around the business plus people who have seen some of these sorts of connections before.  Don’t just ask benefits professionals about benefits and operations managers about operations. Ask bigger questions across a range of stakeholders.

#5.  Falling in love with visualization tools

Don’t do it. Visualization tools are instruments like paint brushes--- amazing in talented hands. Unless you are already skilled, have an in-house data imaginer, or plan to dedicate a lot of time to learning, visualization tools will remain underused. 

Business leaders like the idea of data visualization to unlock mysteries and generate new insights. They become enthralled with the point-and-click ability to see ‘everything!’ And presume that beautiful, futuristic, three-or-four-dimensional graphics will help sell their services and demonstrate value more convincingly that a simple bar or line graph.

That is rarely the case.

When was the last time you heard a busy professional say “what I really want is to spend time learning a brand new, unfamiliar system that has a steep learning curve, using techniques from my least favorite college statistics course.”? Never.

Don’t get me wrong: the ability to manipulate data in ways that reveal underlying structures is a huge advancement with incredible value. These are a nerd’s dream! But they depend on two things: 

1)      Sufficient, high quality data sources

2)      A talented analytic professional who understands both the data and the tool

This applies both to visualization tools for internal use and to those made accessible to customers. Don’t assume you – or your clients – will be able to make use of them easily. I estimate that over 90% of hands-on tools I’ve seen sold to ‘leaders’ have sat unused and/or were assigned to their direct reports to figure them out. Pretty pictures are secondary to having solid skills and systems.

Instead, start with thoughtful integration of data and invest in someone(s) who can help you make sense of it.

Plus, when selecting a data partner, be careful to value the right abilities. If your consultant is prioritizing fancy online search tools (focused on one area, like health) over experience integrating twenty (or thirty) data feeds per client, instruct them to change their criteria.

Photo by Sigmund on Unsplash



Integrate. Integrate everything. Integrate the parts that make your company unique. Integrate across types of expertise. Integrate with a vendor that emphasizes uniqueness and flexibility over standards and volume. Forget the bells and whistles, get a solid data platform and experienced analytic support.



I am a nerd and recovering statistician who helps companies envision their own optimal data integration and run analytic projects.  I have worked on hundreds of data integration projects – with the scars to prove it. 

Wendy D. Lynch's headshot photo.

Wendy D. Lynch, PhD

Wendy Lynch is an experienced sense-maker and data scientist with over 35 years of research experience, primarily in business settings. She has played the role of Analytic Translator for hundreds of companies, from start-ups to Fortune 100 corporations. Her expertise is both in data analytics and effective communication, combining the two into a framework for optimizing the value of analytics in a business setting.  Connect with her through LinkedIn or email.

Newsletter: A Matter of Translation

Making sense where business and analytics meet.
Training professionals to speak both languages.

We won't send spam. Unsubscribe at any time.