The original intent of this blog post was to provide a One Stop Shop for all those digital marketing statistics you find in case studies and sales pages. I wanted to create an encyclopedia for myself that went beyond “here is a statistic” but included the original source of the statistic if possible.
The initial bit of research turned up a few blogs that had done similar things but I am not above replicating other people’s work when I think I can improve upon it. I had this idea of setting up a searchable database and really, truly making it a resource for bloggers and case study writers. If done well, I could update it every year and make it a centerpiece for this blog.
What I found out, almost immediately, is just how antiquated and flawed so much of these statistics we cite are.
The Problems I Found With Digital Marketing Statistics
I am not picking on any one website, any one blogger, any one in particular. These problems exist across the board.
Lack of Proper Citation
While trying to collect all the data points I could, I would come across blog posts like this:
This post is sending every signal that it is fresh – even claiming to be updated on the very day I am writing this blog post. Surely the information in this post has been curated and I can trust it, right?
I do not know if the stats I am about to cite are accurate or not. My issue with them is in how they are cited.
Here is a great example:
This states iPhone users spend more than Android or Windows users on an average order and has a link to https://www.invespcro.com/blog/mobile-commerce/. I remember a similar stat being cited years ago and I really wanted to know more. I clicked through on the link to see what was meant by “average order” and to see if the differences were significant or minor. What I found was that page no longer exists. I tried to do a site search on invespcro.com for this information and could not locate it.
How much trust should I put into this statistic? How many people have grabbed this up, used it in a blog post without digging further?
Out of Date
This issue is the most cringey for me. Using statistical information that is out of date in an industry that is evolving seems crazy to me. Sure, there are some elements that we can assume hold steady over a few years, but take these examples under consideration:
Here is a block of statistics pulled from a page claiming to be updated in June of 2020. These stats are all from Insivia.
I’m going to focus in on #12 there…. “Creating a video landing page has the potential to increase conversions by as much as 80%.” That is quite compelling, especially if I am trying to sell video landing pages. So I click through to Insivia to find out more. Oh, look, the Insivia post was written in 2017. Okay, not a real issue, 3 years isn’t that large of a time difference. But wait. Insivia isn’t actually the source that did the research on this.
Okay, let’s drill down further. I click through on the EyeView link (https://www.eyeviewdigital.com/) and… the site no longer exists. The link itself doesn’t even look like it was going to a page on that domain that continued the information. Let’s persevere and go deeper. Into the Internet Archive we go and like a good Sherman I plug the URL into the Wayback Machine for February 2017, and no statistics on the front page found.
Going to the blog, I can’t find any blog post from that immediate time frame that would clearly indicate it is about conversions which means it probably came from an even earlier blog post… 2016. Making it a 4 year old statistic. Do we believe the nature of video and web conversion has evolved over 4 years? Do we even know what kind of sites that original statistic applied to? Was it all ecommerce? Was it lead gen? There are a lot of questions behind that number that we cannot get answers to.
I won’t dwell on this anymore, but some things to also look out for are misinterpreted statistics. This often happens when a chart or table of data is used as a source and the person reading the table cherry picks a data point from it or doesn’t understand what the chart is depicting. It is easy to focus on outliers and try to make a case around an erroneous bit of data or to not understand the intent of the chart and draw the wrong conclusion.
What is really important to note is when we as blog writers grab these statistics that are being cited by other people who are citing an original work. We need to do our due diligence and dig in further, find that original source. See the date it was created. Understand how the statistic was created, and ask ourselves the hard questions if it is still relevant. If there hasn’t been any updated information on the topic and something from 4 years ago is the best we have, then ethically we need to point that out. We need to indicate that the statistic might not be applicable.
Lies, Damned Lies, and Statistics
Digital marketers are always trying to prove to clients that our services have value. We reach for the best studies available to us. We cite authoritative sources. We do everything we can to build trust.
Using old statistics, statistics that may not be applicable to our client, and statistics that we don’t know what the original source is only harms that effort. Whenever we see something cited as fact with a number attached to it, we need to be skeptical and cautious. We need to be ethical in how we use that information (“One study conducted in 2012 of 3 ecommerce sites selling clothes found an increase of x% conversion when more than one image was used on the product page”). It may not land the same punch but it at least sets an expectation that a) that sample size was very small, and b) the digital marketplace has changed since 2012 so this may not be accurate.
Those of who write about digital marketing in an attempt to sell our services need to be cautious in how we use these statistics. Those of you who are reading case studies and blog posts about amazing results need to be more skeptical. Maybe between the two approaches we can sort out the old, out dated, and just plain bad information and find some nuggets of truth.
What I would love to see is sites that attempt to house these statistics do a better job in curating, dating, and updating their information, especially if they are taking pains to make the page look as fresh as possible.