Skip to content

Why your data is failing you

Written by

Tomas Ausra

On September 4th 2002, with their team tied 11-11 against their opponents Kansas City Royals, the Oakland Athletic’s manager Art Howe decided to sub in first-baseman Scott Hatteberg.

Hatteberg took a second-pitch ball from Royals reliever Jason Grimsley, firing the ball deep into the night to send the Coliseum and its fans into euphoria as this act meant the Oakland Athletics had won 20 games in a row. 

It was history for the Oakland A’s. It was history for Major League Baseball (MLB).

“How can you not be romantic about baseball?” asks Brad Pitt’s Billy Bean in the Academy Award nominated film ‘Moneyball’.

The Oakland A’s accomplished their remarkable success without the budget of any of their top challengers or predecessors. They achieved it by crunching some numbers and having the guts to stand behind them. None of the earlier MLB teams have attempted to do that. It was ballsy. 

It revolutionised baseball. It revolutionised other sports as well.

One could see the admirable story about an underbudgeted team’s success; others might see the power of data and how it is changing the world.

How can you not be romantic about data?

The International Data Corporation predicts that the amount of data in the world will grow from 33 ZB in 2019 to 175ZB by 2025. With so much data at our disposal, it is difficult for us not to dream about our own Moneyball moments.

After all, if we are gathering so much data, surely we can use it to make better predictions, right?

Bullshit.

We like to think that all data is good data, but the sad reality is that it is not.

In 2019, MIT, GroupM, and Melbourne Business School set out to test the accuracy of programmatic data, focusing on the two most commonly used data points in B2C marketing – age and gender.

They found out that the gender metric is accurate only 50% of the time. In other words, marketers would be better off flipping a coin. What’s worse, it turns out that gender is actually the most accurate metric out of all of them. For age, the accuracy drops down to 25%. The more niche you go, the less accurate the data becomes.

The adtech industry is the crème de la crème benefactor and victim of all of this. Billions are being wasted due to ad tech fraud and poor data being used. When Peter Weinberg, Global Lead of LinkedIn B2B Marketing Institute, asked Dr. Augustine Fou, the world’s leading expert on ad fraud, the simple question: ‘‘If the situation is that bad, is it true that digital marketers are reaching the wrong people?’’ Dr. Fou said “No, it’s much worse than that. The problem isn’t that marketers are reaching the wrong people. The problem is that marketers aren’t reaching people at all.”

A whole essay could be written on the adtech industry and the poor practices there but I won’t delve into that black, fraudulent rabbit hole in this article. If you wish to learn more, Dr Fou and the Ad Contrarian, Bob Hoffman, are great sources to follow.

So what does all this bad data mean?

Well, a study from IBM found that poor data quality costs U.S. businesses just a small sum of $3.1 trillion per year. 

recent study by Dun & Bradstreet concluded that 19% of businesses had lost a customer by using inaccurate or incomplete information – a loss even higher in industries where customers have a high lifetime value.

According to Zoominfo, 94% of businesses suspect that their customer data is inaccurate. In B2B sales, that poor data quality is responsible for wasting over 27% of sales team time. 

Research by Royal Mail Data Services revealed that organisations believe inaccurate customer data costs them, on average, 6% of their annual revenues. Perhaps more worryingly, over a third were not sure how much it costs them.

Over a decade ago Marsh, R. found that $611 billion per year is lost in the US by poorly targeted digital marketing campaigns and 33% of projects fail due to poor data. That was in 2005 – the numbers nowadays would have only increased.

As a result, it is no surprise that only 12% of marketers trust the accuracy of their data, according to Forrester

So businesses and marketers know that their data is bad and that it is costing them money, surely they would have this as one of their top priorities to solve for the coming years?

Not even close.

Nielsen surveyed global marketers for the 2019-2020 Nielsen Annual Marketing Report and discovered that audience targeting, ad creative, and audience reach all ranked higher than data quality in their list of priorities. Brands even valued data quality less than agencies did.

All of that bad data is being used to make decisions and measure results.

The problem becomes even more amplified as businesses use all that data to attract investors and prove their appeal. And once it is used, barely anyone will go back and say that the data they used to attract thousands of investments is bullshit.

Imagine you’re the CEO of a fast growing company. You’ve raised millions of dollars as venture capital using reports that show things like website visits, customer acquisition costs etc. If you suddenly find out that your whole business case depends on numbers that are as accurate as a drunk co-worker trying to prove he is a reigning dart champion, what would you do?

So we know the problem is ginormous and everyone is ignoring it. But say someone wanted to do something about it, what could be done?

Firstly, move your data as close to you as possible. Evidenced by the adtech industry, the further away from the publisher you are – the more the chances are that your data will be a product of fraud. You have to move towards first-party data.

Secondly – and I know this might be a hard one – talk to real human beings. When Amazon was just a book retailer and Jeff Bezos wanted to know how he could grow his business, he did something remarkable. He asked them. He sent hundreds of emails and asked what could Amazon sell them to make their lives easier. He received hundreds of different responses and the answer was evident – Amazon should sell all of it.

He kept that close link to the customer by making his email address public and allowing customers to email him directly. Back when Amazon was small, I am sure he read most of those incoming emails. He did not look at second or third-party data regarding what pisses off consumers, he simply allowed people to tell him.

Thirdly – accept that a portion of your data will become outdated every year. MarketingSherpa found that 25-30% of data every year becomes obsolete. If your lead generation campaign converted 10% of prospects last year and now you are targeting that void, is it a surprise that your conversion rates are dropping?

Lastly, marketers should accept a level of uncertainty. When digital marketing exploded over the last two decades and became an integral part of our communication mix, bringing all the data to our systems, we became accustomed to wholeheartedly trusting that data. 

Sometimes marketing works in the world of the unknown – we make future predictions using the best data and research that we can possibly find. Do your best to gather your research, accept a level of uncertainty, and admit that you are making a prediction with a level of uncertainty.

We all want to find our Moneyball moments and I respect marketers that do. But if you want your bat to hit the baseball, then make sure you’ve checked it is not broken and you know where your opponent likes to strike.

Previous article

The creativity crisis

Next article

SD #1: branding, angry customers, Dunning-Kruger, and crypto