Now more than ever, data quality is a serious challenge. Google Analytics expert Brian Clifton shares insights on how to achieve better marketing data quality and governance.

Data quality in Google Analytics

D.D.: You have over two decades of consulting experience with Google Analytics and you're also Google’s first Head of Web Analytics for Europe. How can companies avoid getting bad marketing data in their Google Analytics?

B.C.: Essentially, the problem that almost all organisations face is that an initial good data set-up can go sour very quickly - there are simply too many moving parts. These include not just the website with its constant updates, new product releases, and CMS changes, but also the turnover of digital staff and agencies. Yet bad data looks just like good data - there is rarely a red flag to indicate a problem, and even if there is, it is often buried under the noise.

In short, the solution is to continuously monitor your data quality just as you would monitor your conversion rate. Measuring your marketing data quality is just as important as measuring your KPIs.

The solution is to continuously monitor your data quality just as you would monitor your conversion rate. Measuring your marketing data quality is just as important as measuring your KPIs.

In fact, it's a long-standing problem. Even though the web and its tools have moved on a great deal over that period, the quality of data has hardly changed. This is what I have studied over many years.

Generally speaking, the data quality of enterprise websites is very poor - across all sectors and all regions of the world - including modern digital-only brands that you would have thought would have leap frogged over such problems.

I can say this with certainty because at the core of my work I audit Google Analytics data - it's why I co-founded Verified Data - a tool to automate the process. You can read more about the details of the enterprise study here: verified-data.com/study

How to improve marketing data quality

D.D.: What steps should companies take to ensure accurate campaign data and marketing data quality?

B.C.: Coordinate and centralise your tracking of campaigns as much as possible.

By coordination, I essentially mean training staff and making sure all digital marketers are aligned on why things are tracked in a certain way, what happens when the process is not followed, and how to handle things when there is no obvious tracking solution i.e. figure it out themselves. Apart from an initial workshop (typically a half-day), that could simply be a weekly check-in/review call to keep staff up to date.

​By centralise, I mean one person or team "owns" the campaign tracking. That is, takes responsibility for the above coordination. Not having this centralised approach, even for independent offices in far-off places, is a recipe for disaster. To clarify, other offices should still independently manage their own campaigns, just not do so isolated or in a vacuum. Independence is fine, but measurement needs a coordinated ecosystem for results to be comparable. Otherwise, campaign results and its learnings from Office A cannot be compared with Office B - a common frustration for marketers.

Coordinate and centralise your tracking of campaigns as much as possible to ensure marketing data quality.

Data-Driven vs. Data-Informed

D.D. You've previously discussed being data-driven vs. data-informed. What are the key differences between the two? What approach would you advise adopting, and why?

B.C.: For me, these two phrases represent a difference in time scale, so it's important both approaches are followed and one is not over favoured to any great extent. For example...

DATA-DRIVEN is the more immediate/real-time processing of data. All animals are data-driven – it's called survival. Making a decision in real-time is literally a matter of life or death. For modern humans, we wake up and look at "time data" to make an immediate decision (get up). Then check "weather data" and make the next immediate decision (what to wear), and so on.

DATA-INFORMED is about taking a longer-term view of the accumulated data (only available to humans?). It's the basis of experience and wisdom. It allows us to spot patterns that do not match expectations e.g. discover new things, explore why this is, build a hypothesis, test it by experiment. For example, it looks sunny outside but that does not mean it is warm. Nor does it mean the clothes I choose based on my visual data-driven input are sufficient for the day.

So these terms are quite different – though not mutually exclusive and certainly not one better than the other. If you are not data-driven then you will likely not survive – both figuratively speaking (you miss an opportunity in life) and literally (you crash your car because you ignored the red stop light). If you are not data-informed then the same mistakes are repeated over and over again, and no learning is gained. You will survive, but life would be dull…

How to leverage marketing data effectively

D.D.: In these uncertain times, many companies argue that they don't have enough resources to leverage marketing data effectively. How would you advise them to optimize analytics processes and maximize data use?

B.C.: I would say the pandemic has highlighted just how important and robust the "online" channel is. It's where an organisation needs to continually invest if they are going to survive and prosper moving forward. Finding the right resources i.e. smart humans, has always been difficult even in the good times. And such talent tends to gravitate towards agencies and consultancies where work is more varied with greater creative freedom (analysis is a creative, problem-solving industry).

So finding a good agency with a strong reputation is key. However, ensure that the agency/consultancy works in a way that also delivers "knowledge transfer" to your organization. That way, your internal teams get smarter at the same time and can pick up an increasing amount of work. This allows you to reduce your day-to-day reliance on the agency - giving you the option to save money or work with the agency on more challenging projects.

If you are not data-driven then you will likely not survive – both figuratively speaking (you miss an opportunity in life) and literally (you crash your car because you ignored the red stop light). If you are not data-informed then the same mistakes are repeated over and over again, and no learning is gained. You will survive, but life would be dull…

What's next in marketing analytics

D.D.: If you were to predict the upcoming trend in analytics, what would that be?

B.C.: I will stick my neck out on this one as I have been around banging my privacy drum in the digital website/marketing/tracking business for a while… Essentially I have seen and read a lot of hyperbole as well as a lot of such neigh-sayers. However, something tells me this decade is going to be different. It will be different in terms of how users perceive privacy and as a result how platforms like Facebook, Google et al can make money... So with my crystal ball in hand, I predict the death of personalised ads as we know it by the end of the decade... Hopefully I will still be blogging posting in 2030 to find out...

‍

About Brian Clifton:

With over two decades in digital marketing and Google Analytics consulting, Brian Clifton is recognized internationally as a Google Analytics expert that has helped shape the industry. As Google’s first Head of Web Analytics for Europe (2005-8), Brian built the pan-European team of product specialists. Currently, Brian is the Director of Data Insights at Search Integration, founder of Verified Data, author (x4) and guest lecturer.

‍