Digital Analytics data is never accurate (not 100% accurate anyway). But Digital Analytics is not about data, it is about insights and business intelligence – useful is more important than accurate. Being able to make smarter informed decisions to achieve better results.
Digital Analytics delivers there.
An issue can arise with key conversion metrics, where the actual number is known e.g. transactions, leads, job applications. If the data is too different, too inaccurate, business stakeholders can lose trust and confidence. That is bad.
This workshop focused on this issue. I provided my step by step process to investigate and reduce any discrepancy in transactions or any similar conversion metric.
Audience Questions
What if the discrepancy is less than 5% and management still doesn’t trust the data?
That can be tricky. The first stage is education, explaining why accurate data within Digital Analytics tools is an impossible target. With this closely followed by why 100% accurate data is not required to make smarter decisions and achieve better results.
It is possible that certain people can use the discrepancy as an excuse to dismiss the use of a data set, preferring to use other data they believe is more accurate or defaulting back to opinion & experience to make decisions. There are some minds that are impossible to change. The choice for you then is whether you can work around these people (can’t get 100% accurate data, can’t convince this is not possible/required) or move to a different company where data is more valued.
Is it still possible to get data within 5% accuracy, given the need to get consent to track website visitors?
This is a killer question for the whole Digital Analytics world. The 5% discrepancy target was a common one 5 to 10 years ago. Is it still relevant now given the need to request consent from website visitors to track them? And assuming not, what target should be used these days?
I don’t know what the common benchmark is for visitors refusing consent these days, I hear stories from 20% to 80%. If it is at the 80% level, then the usefulness of the remaining 20% would be called into question.
At one level, ignore the discrepancy. Go through the process outlined in this workshop and resolve any issues that are identified – get the data as accurate as you can.
There are actions you can take with your consent managment platform to improve the acceptance rate, investigate and take these actions.
I am currently reviewing options to record actions of visitors who login and identify themselves, replacing consent with the contract basis for capturing data. A key aspect would be to measure behaviour but not track visitors, doing the minimum possible to provide any business data needed. Updates may be provided in the future on this (get in touch if you want to discuss).
Any more questions?
Let me know your thoughts and any questions on this process I recommend. Are there any actions you use to identify or resolve discrepancies that I didn’t list here? If so, please share them in the comments.
If you have found a discrepancy and need support to identify and fix the issues, please get in touch. Get trust in your data so your business can make the best decisions to achieve success.
