Perhaps as an analyst or digital marketer you’ve had that awkward situation where an analytics stakeholder looks at you askance across the board room table and says, “I don’t think these numbers are right.” And then it’s up to you to defend the quality of your data.
When numbers don’t confirm their intuition, stakeholders get suspicious. While your reaction may be one of frustration, sometimes numbers are wrong, and it’s the stakeholder’s intuition that holds true.
Still, that data is your livelihood. It’s your bread and butter. You want to get it right.
So what can you do to help stakeholders not only trust the data but use it in their decision-making processes?
Data Needs Validation
In order to allay the fear that data is incorrect, you need to show stakeholders that it has been validated.
I remember a college economics class where data was king. Every grueling assignment was full of figures, numbers, formulas, assumptions, variables and constants.
One day our seasoned professor asked us how we could tell if a number was wrong. Our ears perked up, thinking that he was about to reveal the secret to the economics universe—and our homework.
He answered his own question: “You can tell a number is wrong if it hasn’t been checked.”
It was a good lesson that proved true time after time—no number is correct that hasn’t been checked and rechecked.
And then checked again.
Now, you can’t physically watch every visit to the site to make sure all the pageviews are accounted for. You can’t call up visitors and ask them to tell you which pages they went to. Yet you hope with all your data-loving heart that the numbers are right.
But how can you be sure?
There needs to be a process of verification so your analytics stakeholders have confidence in your reports. Data validation is like a stamp of approval. It means your numbers are independently verified. Without this seal of validity, why should anybody believe that the numbers match reality?
Data Validation the Wrong Way
Some analysts and marketers will set up a second web analytics implementation as a form of data validation. They lean back in their chairs and think to themselves that if anything goes wrong with their main analytics tool, their secondary tool will somehow notify them of the issue.
Initially this sounds like a logical solution, since hypothetically the number of visitors from one tool should match the other, and that by comparing the metrics of one to the other you will be able to see if one is incorrect.
Sorry, but it doesn’t work that way.
Secondary tools are fantastic analytics solutions and may be used to easily integrate advertising tools, but they were never meant to be validators.
First of all, analytics solutions vary in the way that they measure metrics, and are therefore not equivalent to each other. A bounce rate in Google Analytics, for example, is not the same as a bounce rate in Adobe Analytics.
Second, how do you know which tool is correct to validate the other against? If you already know that one data set is correct, why are you validating at all?
Most digital marketers I’ve seen don’t try to actually validate every variable with the secondary tool because it is just too much work. But if you aren’t duplicating every variable in each tool, what’s the point of believing the secondary analytics is a validation tool? It doesn’t make sense.
If you still haven’t grasped the futility of doubling-up your analytics, let me suggest another argument against this practice: Trying to align one analytics tool to match the metrics on the other tool requires a lot of resources and man-hours.
Do you really want to allocate tons of resources to the pursuit of meaningless congruency between two tools?
That may not be the best way to gain the confidence of your stakeholders.
Data Validation Done Right
A secondary tool can’t:
- Count how many page views the other tool has collected
- Show you pages that have missing tags
- Show where you risk over- or under-counting metrics
- Alert you when a step in the conversion funnel breaks
A third-party, automated validation tool can.
Plus, it has the advantage of being an independent technology that browses your site just as a real person would.
ObservePoint’s Data Quality Assurance™ solutions make it possible to validate data without the hassle of manual validation that eats up resources. It is an automated technology built specifically for data validation and is what you need to ensure high data quality and to gain credibility in your organization.
It will give you the confidence to look back across the table and say, “Yes, those numbers are correct.”
About the AuthorLinkedIn More Content by Matthew Maddox