Your data tells lies – digital marketing has a dirty little secret
Is your data doing more damage than good? According to James Lawrence, marketing data in most companies has serious issues and limitations, and if used poorly, will deliver unexpected results.
As an agency owner, one of the scariest things I hear from a client is that they intend to make all of their decisions based on data.
Why might this be concerning? Because I know the dirty little secret of the digital marketing world… that marketing data in most companies has serious issues and limitations, and if used poorly, will deliver unexpected results. At times, it will do as much harm as good.
Let’s face it, even in digital, many valuable activities taken by prospects do not generate a click, form submission or sale. Nevertheless they help prospects become customers further down the line.
In addition to being limited in telling us what’s happened in the past, data is also limited in its ability to tell us what to do next. Often when deciding on the best course of action, a combination of data and our experience and intuitive skills as marketers is what’s needed.
In a nutshell, that’s why decision making that is 100% based on data genuinely scares me.
On top of this, many campaigns we see are suffering from one or more of the following issues.
- Vanity metrics that lack business insights
It’s always tempting to focus on ‘vanity metrics’ while ignoring the more challenging metrics that will actually drive future revenue. Before you focus on a metric, you need to ask yourself if you’ll be able to make a business decision based on what the metric tells you. If you cannot answer this question, you’re probably focusing on the wrong thing. The worst offenders here are often the most senior people in an organisation. As a marketer, you should be prepared to defend the value of metrics others might not initially be comfortable with.
- The data being collected lacks meaning and accuracy
Data must also be accurate for it to be meaningful. When our data specialists audit new client campaigns, they often find issues with how data is being collected. Roughly half of the campaigns our specialists analyse have significant configuration or tracking issues — Google Analytics has been incorrectly installed, it’s double-counting key metrics, it’s not picking up goal completions or it’s tracking goals that are largely meaningless.
Scarier still, we often find: the larger the company, the larger the issues with data collection and accuracy. When was the last time your data collection was audited from the ground up?
- Marketing data is often not regularly reviewed
Most companies use Google Analytics or a similar packages on their website. Data is also collected in Facebook, Google Ads and many other platforms. The first mistake many marketers make when it comes to data is in not actually reviewing this data on a regular basis. If it’s not analysed, all the data in the world means nothing. Honestly, when was the last time you checked your analytics package in detail?
- Attribution is not understood or valued
There are many marketers, and even more executives who find themselves stumped when it comes to a discussion about attribution. Trust me when I say you need to understand this valuable area and how it affects everything to do with your data.
Attribution, simply explained, is how you credit separate channels, or touchpoints within those channels, for the role they played in a prospect’s path to purchase or goal completion.
If people visited your site a single time and then either converted or left for good, attribution would be simple. The thing is, this is almost certainly not how your buyer journey works. What happens in the real world, where most paths to purchase are complex and involve multiple visits on possibly multiple devices? How do you determine which channel deserves credit? Is it the one that drove them to you in the first place (this is known as first-click attribution)? Or is it the channel that delivered them to you when they finally reached the goal (known as last-click attribution)?
Alternatively, does it make more sense for you to equally value all the channels that drove a person to your site prior to their converting (this is called linear attribution)? These are only three of many ways to attribute credit to different marketing channels whenever a goal is reached.
If you’re not clear on the basics of attribution, then I’d strongly recommend you re-read the last couple of paragraphs.
The truth is attribution is a critical part of making decisions when using marketing data, but you need to know what attribution means for you, keeping the full buyer journey in perspective. Chances are that your analytics platform is defaulting to a last-click attribution model. For most people this is very bad and can result in poor data-driven decision making.
Ask yourself: how complex is your buyer journey? How likely is it this model undervalues the earlier stages in the buyer journey and overvalues the final stage? How is this affecting the way you judge the performance of each channel? What impact is this having on how you set budgets and select future channels?
- Too much focus on quantitative data and no focus on qualitative
It is critical for a smarter marketer to go beyond the quantitative data collected by their analytics platform and start collecting qualitative data to help their decision making.
Qualitative data is typically non-numeric and largely based on what you can learn from what people say and do. That’s right, it’s about speaking to real people and looking at how individuals interact with ads, landing pages, websites and more. It’s about survey data, heatmaps and more.
- Jumping to conclusions not correctly supported by the data
It is very easy to understate the complexity of data and overstate the insights the average business can receive from the relatively small amounts of data they have access to. I use the term ‘small’ to refer to business of all sizes, as most of us overestimate how much data we have and underestimate how much data is needed to be truly meaningful.
A lot of articles you’ll read talk about how powerful data is, but those articles often base their conclusions on large companies with highly sophisticated data sets. Google’s machine learning within Google Ads is constantly evolving and improving campaign outcomes, but many small and medium-sized businesses don’t benefit from this as they don’t have the volume of data required for machine learning to be able to make accurate decisions. The more complex the question you want data to answer, typically, the bigger the dataset you’re going to need.
Then there are the many valuable actions never measured that I referred to earlier in this article. A prospect might see an ad and be influenced by it, but not click on it. Someone else might receive and read an email from you and then later manually type in your website or search on your brand name.
It’s impossible to measure the value from these sorts of activities and that’s completely okay. In traditional marketing, hardly anything was trackable and yet value was created through marketing activities. In digital it’s fine that not everything can be attributed perfectly.
Data can deliver powerful insights when used properly. But it can also mislead or tell outright lies, especially when you push it to answer impossible questions.
Where to from here?
Firstly, and most importantly, don’t burn your Google Analytics logins and storm the barricades to get to the people who created your potentially flawed marketing reports. Instead I’d suggest you take a more mature stance when it comes to data in general.
It may not have been clear from this article, but I love marketing data and use it to make many excellent decisions. That’s where I want you to get to as well.
As a first step, take some time to sit down with your team and agree on a small number of metrics that will genuinely give insight into how your marketing activities are contributing to company profits. Then work backwards to develop and improve campaigns to get those metrics moving in the right direction. But don’t forget, consistency is key, and to start with, keep it simple and sceptical.
I’d also suggest you have someone audit how data is collected on your web properties. Is the installation correct? Are you tracking the right sorts of actions? Will this information help you make valuable business decisions?
Finally, keep asking yourself, how much faith should you have in your marketing data?
James Lawrence is cofounder of Rocket and author of Smarter Marketer: 11 Golden Rules to Help In-house Marketers Thrive in an Ever-Changing Digital World
- Brain trust: how would you do data differently? »
- I dream of (data) genie – power, pace and paradox »
- How the data ethics discussion is impacting Australia’s largest analytics business »
Image credit:Kristina Flour