Analytics Predictions 2019
Now and then I tease myself by trying to predict the future, and so in preparation for this years’ attempt I found myself reading the recent Forrester Prediction 2019: Business Insights report.
This part stands out:
57% of global data and analytics decision makers are still at the early stages of their insights-driven business. Only 8% demonstrate advanced insights-driven competencies.
Meaning, that despite the hype of recent years around machine learning and artificial intelligence, even the biggest companies struggle to extract value from their data at scale and drive impactful business results.
Mulling over this insight, I wanted to add my thoughts for why organisations are in such a low state of analytics maturity. Based on working with senior managers in large brand-leading organisations, these are my top 3 reasons:
- A Lack of Staff
For example, having smart personnel capable of driving a “data informed” culture (rather than blindly data driven) i.e. to achieve a greater level of analytics maturity.
- A Lack of Budget
I am often surprised to find that even when an organisation has invested in an analytics team, at how small that team is compared to the scale of the task. Often key people are not dedicated to analysis, and with a modest budget (if any) for 3rd-party help expertise.
- A Lack of Trust in the Data
I consider this the BIGGEST hurdle to what holds leadership back from investing in the above. That is, the data accumulated simply “smells bad”.
Would you invest in a bad smell…?
For web analytics data, the last point – a lack of trust in the data, is the biggest hurdle facing organisations. My question is therefore a rhetorical one – of course no one wants to take a risk on analysing poor quality data! And this is a widespread issue. In my recent Enterprise Study of Google Analytics Implementations I found a depressing state of data quality that organisation are not even aware of. For example:
Issue A – Poor data quality in all tracking areas
When ranking a website by its overall Data Quality Index (a weighted score from 0-100 where 100 = the best possible data quality), the average is only 35.7. Moreover, only 12% of sites have a score above 50 – something I insist upon exceeding before analysing data.
Issue B – Personal Data Captured
Incredibly, one in 5 websites have a PII issue i.e. personal data being collected in Google Analytics. Often this is by mistake – names and email addresses hoovered up in URLs and page titles by GA. However, in can also be deliberate captured as an event action/label, custom dimension, affiliation code or other variable.
Issue C – Even the basics can go wrong
Half of all sites have issue with deploying the Google Analytics tracking code. That is, the basic code that tracks visits and pageview data. With holes in a GATC deployment, visit counts can be duplicated or lost completely, as well as producing unreliable attribution results.
Issue D – Poor visitor segmentation
Segmentation is a key requirement to be able to perform any kind of in-depth analysis of data. Yet, visitor segmentation was by far the most poorly understood/implemented feature of Google Analytics with only 7% of websites getting segmentation right. By default GA has some great default segmentation tools. However, these are at the session level – they do not tell you about your users i.e. real people. Read my definition of what is tested with respect to visitor segmentation.
And so on… See the post: Enterprise Study of Google Analytics Implementations for further details of the 15 key areas of data quality audited and summarised.
When it comes to web analytics data, there is such a lack of trust in the underlying data, that management avoid the serious risk/investment required to make it happen. These reasons are why I think only 8% of enterprises in the Forrester report demonstrate advanced insights-driven competencies.
Data Noise Obscures the Signal
(or, how such issues fly below the radar)
I use the “bad smell” metaphor to summarise that senior management are uncomfortable with data quality when it comes from anonymous sources. That is, the vast majority of web analytics data is not from customers, but from anonymous prospects. Anecdotally, senior managers are aware that numbers just don’t add up, contradict other sources, or simply cannot be explained. However, it’s difficult for anyone in an organisation to put a finger on exactly what the issue is. There are literally so many wheels in motion at any one time with online visitors, that the data noise obscures the signal.
The Solution – Stress Test Your Google Analytics Data
With so much going on, so many data points, and the time pressures that the online world brings to a business, organisations need a easy way to stress test their data quality. The traditional method has been to manually audit your data and account setup looking for issues with tracking code, configuration and use of advanced features. I discuss a methodical way of doing this in detail in my last book (Chapter 4).
However, the trouble with manual auditing is that it is time-consuming. For example, for an enterprise website I could spend 20-30 hours preparing an audit! And it’s not something that can be off-loaded to an intern/office junior. Auditing a GA setup is a forensic process requiring expertise and experience. Another issue with manual auditing is its fragility – even experts make mistakes, especially when you are looking for the proverbial needle in a haystack. When it comes to web analytics data, the needle looks just like the hay, and the haystack is constantly growing and shifting.
Hence an automated auditing approach…
So what started off as a project to make my job easier has now become a fully featured enterprise cloud tool. Verified Data does all the heavy lifting of assessing and verifying your Google Analytics data. It uses hybrid technology to avoid human frailties – combining a website crawler to discover content that should be tracked, with forensic assessment (and some artificial intelligence) of the data collected. It also real-time monitors governance issues, such as GDPR compliance.
And my 2019 analytics prediction…
The point of this post was to provide a background story on what I predict for web analytics in 2019… The past 12 months has been a turning point for data governance i.e. GDPR. The greater emphasis and responsibility now placed on gatherers of data will inevitably lead to organisations scrutinising their data quality. Hence I am expecting the following question to rise to board level for web analytics: “Can this data be verified?“. And if not, go and verify its quality.
That’s good for all data users, and hopefully for Verified Data 🙂