Why Do We Frequently Question Data But Not Assumptions?

United States President Donald Trump recently stated he felt the contributions that data made to Obama’s successful presidential campaigns were overrated. Meanwhile, it’s reported that Hillary Clinton assembled a data team that was three times larger than Obama’s impressive 2012 operation.

Despite the increasing investments that companies have made in analytics tools, many people still align more with Trump’s sentiment and don’t want to rely too heavily on data. For example, while analytics was in the top three spending categories for US banks, only 20% of banking executives wanted their organizations to be highly data-driven. For some individuals, it can be unnerving to trust data that is difficult to fully understand or which doesn’t align naturally with their intuition.

Frequently, this fear leads people to incessantly question the validity of their data. Many business owners and managers are nervous that erroneous or incomplete data will lead to bad business decisions. For example, film manufacturer Kodak forecasted in the late 1990s there would only be 400,000 digital cameras on the market (outside of Japan)—however, the estimate ended up being off by over 4 million units. A more accurate estimate might have helped Kodak to better anticipate the growth of digital photography and chart a different course where it could have avoided Chapter 11 bankruptcy.

A healthy dose of skepticism is sometimes necessary when it comes to your company’s data. It is never going to be perfect and is most likely peppered with various discrepancies. As the value of your data increases, it needs to be managed over time to ensure it’s as consistent, reliable and useful as possible. Many data-driven companies have instituted data governanceprograms to avoid or lessen serious data quality issues.

However, why don’t we also see a healthy distrust of business assumptions? Why is there no such thing as assumptions governance? You may have heard the great quote attributed to Mark Twain, “It ain’t what you don’t know that gets you into trouble. It’s what you know for sure that just ain’t so.” Even though there’s no evidence Twain ever used those words, I still like the quote despite its suspect origins.

Assumptions seem to get a free pass when data has to fight for every little shred of respect and acceptance. Poor assumptions can be equally or more damaging to businesses than bad data, but they receive far less scrutiny than data does in most situations. Faulty assumptions about film photography probably did more damage to Kodak than bad data ever did. Kodak actually invented the world’s first consumer digital camera in 1975but refused to launch it due to fears of how it would affect its lucrative film business.

To say you should never assume anything is unrealistic and often unproductive. A list of financial and business assumptions is a key section in any entrepreneur’s business plan. In a dynamic, fast-paced environments, it can be hard to operate without assumptions. Both startups and established companies need to make them when all the facts they need to form decisions aren’t readily available. It may be too costly, time-intensive, or difficult to obtain the data you need so you formulate assumptions—explicitly or implicitly—to move forward.

However, as your organization makes assumptions—knowingly or unknowingly—it must treat them as unverified placeholders until they can eventually be substantiated with data. Too often underlying assumptions are never revisited or validated, and they end up blending into the essence of how you operate. Because some assumptions are accurate or true, they don’t necessarily lead to problems. However, assumptions that “just ain’t so” can leave you exposed and vulnerable to costly mistakes.

Tim Wilson of Analytics Demystified shared with me an example of how assumptions almost led astray a CPG company’s marketing efforts. The marketing team was running a four-month campaign to drive trials of a new product, which included both samples and relatively high-value coupons. They were nervous about deal sites becoming aware of the campaign, which could attract low-value deal seekers who wouldn’t translate into the desired high-value customers they were targeting.

As a result, they worked with their agency put in place a number of technical measures to restrict who could gain access to the campaign. However, for a two-day window near the end of the campaign a technical glitch led to several Facebook FB -0.48% pages devoted to finding and posting coupons and deals discovering the campaign offers. The company was disappointed to learn almost a third of the registrations generated by the four-month campaign came during this two-day period.

An analyst decided to test the theory that these deal seekers were indeed low-value customers. A key part of the campaign was to have consumers respond to a follow-up email to rate the product two weeks later. As there was no additional reward for doing so, it was viewed as unlikely that these deal-seeking customers would respond after they had already secured their discounts.

However, to their surprise the rate at which these users responded to the email and rated the product was twice that of any other group of consumers. Only by chance was the marketing team able to avert a major misstep that would have cost them a key segment of highly-engaged, valuable customers.

Both facts and assumptions are involved in decision making. In military briefings, they are explicitly stated upfront so the commander has a clear understanding of the current situation. While facts are often discussed in business meetings, assumptions receive much less attention or scrutiny. Business owners and managers should consider the following three steps to better govern assumptions:

1. Do an inventory of what’s true. When you’re working hard to move your business forward, you may be surprised by how many assumptions are actually influencing your decision making. For the key principles guiding your decisions, ask yourself how many you think are true (assumptions) and how many you know to be true (facts). For example, you may believe your customers would be very interested in a bundled offering, but without supporting data it may just be wishful thinking. In addition, your facts may actually be just assumptions if your data is outdated or questionable. Don’t let assumptions masquerade as facts.

2. Weigh the impact of your assumptions. Once you have a better sense for what assumptions you’re leaning on, evaluate which ones would upset the apple cart if they weren’t true. For example, if you assume that a major competitor is ignoring your corner of the market, evaluate how your plans would change if they were to launch a competing product in your area. Being aware of how dependent your strategy is on certain assumptions being true will put you in a better position to prepare and react quickly if a key assumption suddenly unravels.

3. Challenge your key assumptions. Rather than waiting for an assumption to potentially fail, it’s better to be proactive and validate key assumptions before they put your success at risk. Disproving a key assumption is incredibly valuable—like discovering a tear in your parachute before you skydive. In some cases, you can run a small test to verify a key assumption before a full-scale launch. While it may delay your launch slightly, it will either give you the data to move forward with confidence in your strategy or valuable insights into what adjustments may be necessary to be successful.

Being data-driven isn’t just about how effectively you use data but also how you manage your assumptions and biases. Interestingly, your data can be 100% sound, but it can still lead you astray if it’s based on faulty assumptions. Instead of being fearful of data, view data as a way to separate wrong assumptions from right ones. It requires discipline to trust your data and want to prove yourself wrong.

As you become more aware of what’s influencing your decision making and seek to replace assumptions with facts, you’ll end up making better decisions. As the philosopher William of Ockham stated, “The explanation requiring the fewest assumptions is most likely to be correct.” This type of thinking has separated data-driven companies like Google GOOGL +21.78%, Amazon, LinkedIn and Netflix from their competitors, and it can help yours as well.