Common excuses preventing you from doing analytics at your company
Here is how you can dispense them
After writing blog posts and working on Humanlytics for the past 3 years, I began to realize that the biggest blockade preventing businesses from adopting analytics at their organization is neither lack of technology, nor talents, but rather the stubborn nature of the human mind.
Let’s be honest, our minds (and the minds of our co-workers or superiors) are lazy, rigid, and fallible, making up excuses in our mind that are very often completely false to prevent us from doing things that we SHOULD do at our business.
So let’s talk about those excuses today, and see why almost all of them are exactly what they are — excuses.
I hope you can use the result of this exercise today to dispel barriers preventing either you or your boss from adopting analytics at your company and start reaping the benefit of a more analytics-driven management method.
Excuse 1: I don’t have enough data at my company to do analytics
My immediate reaction upon hearing this excuse is — if you already know you don’t have enough data, then better go start collecting today!
But let’s slow down a little bit because I can almost certainly guarantee you that you already have enough data at your organization to start doing some analytics tasks.
Very frequently, our mind will force us to focus on things that we DO NOT have, instead of what we DO have, and this principle cannot be truer when comes to data at your organization.
If you are a B2C business, you have to in some way maintain a dataset of your revenue from different products either by day or by week (the POS system); and if you happen to be a B2B business, you have to have some way of organizing your list of customers and your engagements with them (the CRM system).
If you are selling the aforementioned product online, you most likely will have at least three sources of data — the platform you are using to run advertisements, the analytics platform for your website (such as Google Analytics), and your e-commerce store.
In fact, for most businesses that I have talked to, their issues are having TOO MUCH data at their organization, instead of too little — and the data they already have can already provide a significant amount of value to their business.
You may ask the question “but Bill, I know I have data at my organization, but those data are simply not complete enough for my analytics to begin. For example, Google Analytics is horrible at tracking event based funnels.”
The problem with this “I don’t have enough data” mindset is that you will NEVER have the complete dataset you desire to start analytics.
Even the executives of major brands such as Macy’s or Amazon will have the same complaint in their analytics efforts. This is because there are always more data that can be collected, and more analyses that can be done — and you should not make this an excuse for you to not do ANY analytics.
Here is my tip to dispel this “lack of data” mindset: Instead of focusing on what data you are missing, and what analyses CANNOT be done with the data you have, focus on what you CAN do with the data you have.
For example, while you might not be able to do a continuous funnel analysis using Google Analytics, you can analyze the statistics of your key funnel pages independently, and create a “pseudo funnel” that achieves 80% of the analytic power of a regular, continuous funnel.
Being able to work with data constraints is what distinguishes good analysts and the great ones — you should (and will) be the great one.
Excuse 2: Our data is not clean enough
Connected with the previous point, even when data are available at your organization, I hear a lot of analysts and business owners complain that the data they have are not “clean enough” for their analysis.
Indeed, data collected by tools such as Google Analytics often run into the issues of broken sessions (resulting in two sessions recorded instead of one), inaccurate attribution, missing sessions (due to cookie disabling), etc.
To bypass this excuse, we need to fight the desire for perfection that is intrinsic to all of us and be conscious of the true impact those imperfections have on our data.
Let’s face it, the internet is messy, and the technologies surrounding it are designed and developed by imperfect humans.
What this means is that there will ALWAYS be cases and ways your data will be tracked inaccurately — however, most of those inaccuracies does not matter.
One important thing we need to realize here is that business analytics is not rocket science and does NOT require your data to be 100% precise, so to conduct analytics, we are not looking for the most accurate data, but rather the most consistent and systematically stable data so those data inaccuracies are not misidentified as business signals in our analyses.
To be consistent, our data must be always inaccurate in the same way, predictably.
For example, if you have identified a “broken session” error in Google Analytics that is causing you to have on average 10% sessions recorded in Google Analytics.
As long as this error mentioned above is always around 10% of the sessions (low percentage), and happens every single day (consistency) when we are tracking data using Google Analytics, we can safely ignore this data inaccuracy and proceed with analytics tasks since this consistent inaccuracy will not impact business signal generation for most analysis that we will run.
Now let’s move onto the second point, which is systematically stability.
A systematically stable data is inaccurate across all dimensions equally.
Let’s take the “broken session” error again as an example.
If the “broken session” error only happens to users using the Google Chrome browser, this error needs to be fixed before we perform certain dimension-related analyses such as identifying the single most important browser to optimize for.
This is because inaccuracies happening only on one dimension attribute will artificially inflate or deflate numbers of that specific attribute only, sending misinformation to our models that will eventually cause the model to be inaccurate.
However, even with such dimension specific error, most analyses can still be conducted as long as they are not specifically related to the dimension in question — so even those errors are only preventing you from running SOME analyses, rather than all of them.
Finally, let’s talk about platform consistency.
We don’t have enough time to do analytics at our company
We don’t have enough expertise to do analytics at our company
Resources are available, worst comes to worst hire someone to do it for you.
This article was produced by Humanlytics. Looking for more content just like this? Check us out on Twitter and Medium, and join our Analytics for Humans Facebook community to discuss more ideas and topics like this!
For more reading: