Could Data-Driven Decisions Be the Death of Your Company?

Data must be clean, comprehensive and in context before it can be trusted and properly used.

Photo of Justin Spencer

Justin Spencer

Senior Director, Data & Intelligence

Numbers do lie. It can sometimes be intentional, but often it comes from other factors. Even the sincerest efforts in collecting, understanding, analyzing, and acting on data can have dire consequences – as in your data-driven company could be driving straight off a cliff.

According to a recent report from HFS Research, only 5% of CEOs completely trust their data. They are right to feel this way – in my 15 years of experience helping companies use data to ask smart questions and make sound decisions, I find that many leaders are relying on data that is incomplete and unreliable. Flawed data hinders an organization’s ability to fully understand its audience, jeopardizing their potential to retain their current audience and court new customers. But when we can combine reliable data with our intuitive understanding and experience, we hit the sweet spot. It’s science meeting art.

Many companies tout being “data-driven,” but at Moore, we go beyond this to ensure the process we use for our clients is data-informed. This means data doesn’t only answer questions; it prompts them. Gathering data and reading reports aren’t enough. You have to act on the data. And in order to act properly, you have to trust it.

Confusing Data is Bad Data

When you look at your prospect and customer data, you need to know what it means. Even further, anyone who looks at the data should know what it means. About 10 years ago, my mentor asked me, “If everyone on your team dies tomorrow, could the next person look at their data and pick it up immediately?” (It’s a bit disconcerting that we had this conversation in a prop plane flying over the mountainous Pacific Northwest, but that’s a different story.) I carry that same mentality with me now, albeit a bit less morbid.

Data-gathering applications such as Google Analytics are designed to cast a wide net across billions of websites, so the one-size-fits-all approach quickly turns into one-size-fits-none.

If you logged into your Google Analytics data right now, could you immediately recognize “PromotionA_3_2019” or “summer-special-announcement”? What is the difference between “/contact-us” and “/contact-us/” in your Pages report? Why is “Paid Social” showing up as “(not set)”?

When analyzing data leads to confusion, many executives choose to abandon their data altogether. According to a 2019 Deloitte survey, 67% of US executives are not comfortable assessing or using their own data. You're not alone, and Moore is here to help.

Inaccurate Data is Bad Data

Data inaccuracies come in many forms. For your website data, it typically comes from over-counting your website visitors due to improperly implementing Google Analytics. There are a few telltale symptoms of bad Google Analytics implementations that we look out for, including:

  • Incredibly low Average Bounce Rate (it’s healthy to be around 30%-40%)
  • Low Average Time on Page (you might see “<0:00:00” in your data)
  • Low Session Duration

If your website goes from www.example.com to subdomain.example.com, that’s an entirely new set of problems. Out-of-the-box Google Analytics will treat this as a new website and completely skew your data.

Beyond website data, research company Syniti has analyzed thousands of projects and determined that almost 90% of inventory and supply chain problems are rooted in bad data, costing millions in inventory and personnel costs.

If this describes your data, you might have a problem with data integrity. It’s time for an audit.

Data Integrity

The foundation of any relationship is trust. This axiom is true for your data, as well. Ask yourself: Is my data trustworthy? At Moore, we are continually asking this question and we take the time to audit the data our clients have collected to ensure it meets our CORE criteria:

  • Is the data comprehensive? If you are making business decisions from partial data, it leads to bad choices.
  • Is the data organized? If you can’t efficiently access website, CMS or supply chain data, you waste time, resources, and opportunities.
  • Is the data readable? If you can’t immediately understand your data, the inefficiencies and misunderstandings lead to bad business decisions.
  • Is the data executable? If you can’t actually learn or act on the data, there’s no reason to collect it. Data is full of potential energy, but only if you act on what you’ve learned.

Once data is trustworthy, it delivers insights that must be executed across the business. For example, a sales manager could compare Instagram likes with warehouse inventory and direct his team on the best products to push. In a slightly more nuanced example, at Moore, we’ve been able to analyze paid media performance and apply those learnings to our Public Affairs efforts. Our audience’s ad clickshare shows us the most active cities, time-of-day they are online and language that drives action. We then use that information to plan media events and broadcast promotions.

Data doesn't work if you keep it to yourself. Moore will help you use it strategically to take action.

So, how can you ever trust your data?

The ability to trust and use your data will become increasingly important. Only 23% of Fortune 5000 companies employ a consistent and policy-aligned data strategy.

Moore incorporates a data strategy for all of our clients, from research and implementation to optimization and post-campaign analysis. Every decision we make is not data-driven, it is data-informed. Data-driven organizations recognize the importance of data but often fall short on using data to create meaningful action. Our data-informed mindset focuses on the ability to gather the data, ensure its integrity, and contextualize it with our industry experience to craft a well-rounded campaign that resonates deeply.

Justin Spencer is Senior Director of Data and Intelligence at Moore. To schedule an audit of your organization's marketing data, click here.

Fin