Been almost a year since my article on the E3 method was published. A big thank you to everyone for their support! 🙂
I wasn’t able to do justice to all the appreciation: certain unfortunate events kept me away.
But I’m happy to say that the very first E3 Case Study is finally here.
For those who are new, the E3 Method of Website Conversion Rate Optimization stands for:
- Explore – what has worked for others.
- Examine – how similar elements are performing in your case.
- Extrapolate – the data to find viable conditions for probable success and further testing.
The only thing new about the E3 Method is the name I’ve coined for it.
It works wonderfully but is generally so marginalized that I thought it deserved a name of its own.
It is something we’ve all tried at least once, but most likely without a clear understanding of its potential.
E3 is intended as a starting point for CRO with minimal or zero investment.
It is not a replacement for the usual data based (and big budget) approach.
What this case study is about
When I stumbled upon the potential of the E3 method, I was working on a CRO article for a client. I could think of its application only in CRO projects. However …
As I began drafting the present article, it was clear that E3 could be applied to SEO as well.
There are two parts to this case study.
PART I – E3 CRO
Enhance rank through better visitor engagement and,
Ensure rank in the foreseeable future (one year and counting).
Less than 30 minutes.
Actual work done:
Insertion of four lines of text in the original content.
( Can’t wait to see the results? Click Here!)
PART II – E3 SEO
Rank an article – when Brian Dean’s Skyscraper Method fails to work .
Since E3 was originally about CRO, I’ll come to the SEO part of the case study in a separate article after we take a look at the CRO application.
PART I: E3 CRO
Enhancing rank through better visitor engagement
<Explore – what has worked for others. >
<Examine – how similar elements are performing in your case. >
An article I had ranked through E3 SEO had fallen in rank. It was moving up and down between the 6th and 7th result on Google SERP.
I took the simplest route – the E3 route, and checked out the competition.
Now, this ‘checking out’ part needs to have some kind of predetermined direction.
If you look at something without knowing what it is you’re looking for, chances are, you’ll notice nothing.
So, before anything else, I needed a hypothesis.
This is what I knew already (and confirmed, as an obvious starting point):
- None of the other nine results had better content.
Ranking depends upon a number of different factors not all of which can possibly be under one’s control. I did not take things like DA or backlinks into account because those are not factors one can optimize or change overnight.
For an article that was already ranking as mine was …
It is conceivable that Google may have noticed a distinct decline in user engagement and decided to penalize the article.
So user engagement it must have been, I figured.
‘I figured’ is not a terribly technical term, yes. But then, there’s nothing technical about the E3 method. Anyone can apply the principles and expect a reasonable degree of success.
Now that I had my hypothesis, it was a simple matter of looking through the results on SERP to understand how visitors might prefer them to mine.
Thankfully, it was easy to find:
- The article just ahead of mine had something that my content did not.
It had listed several brands (like I had) but before the list began, there was a section on the ‘top picks’ with three lines of text for the best three brands.
A visitor could, if they so wished, click on any of those brands’ names and go to the relevant section of the article.
Or so I thought.
I found that I liked the idea (as a visitor). I could always browse the other brands later but, I was being given a very clear road map on how to proceed.
In other words, …
I liked that I was given the opportunity to choose from three instead of, say, nine.
There was something else this site (and some of the other sites) had which I do not particularly take to:
- The Pros and Cons way of describing a product.
Since I had two things to choose from (making both changes at once wouldn’t tell me much about the difference between the control and the test pages), I went with the one I liked.
Later, I did some digging and found that there was actual research that could explain why limiting choices is a good move.
And, that there’s a whole lot of material on how trusting one’s instincts while making a decision is mostly a good thing. Google for Dr. David G. Myers and Dr. Judith Orloff if you’re interested.
But something interesting happened when I clicked on one of the three links at that site (just to more fully experience this good feeling I was having) …
I was NOT led inside the article to the relevant section that described to me more of the product.
Instead, I was taken to the seller’s site directly.
This I didn’t like for two reasons:
- I had come to the site to make a purchase decision, not to let the site decide for me which ‘top pick’ I should buy
- This would drive visitors off my site faster and cause visitor engagement to actually decline.
Admittedly, this new discovery makes it seem like my E3 driven conclusion was wrong, after all.
But not really …
There was still a chance, however slim, that my initial assumption was correct, and that was because –
- The links, when clicked, opened new tabs.
Which means, visitors were being given the opportunity of choosing from the small list of three as well as the option to stay on the site and browse further.
Agreed. Nothing about this is as impressive as a proper CRO case study.
But tell me, …
Would you or would you not invest less than thirty minutes of your time to attempt conversion optimization before calling in the big guns and spending a bundle on them?
That’s what I thought.
Moving on …
<Extrapolate – the data to find viable conditions for probable success and further testing.>
I inserted these four lines (not counting the one in brackets) into the content before the list (and the comparison chart for the brands – more on that later):
The three links in blue led to the relevant sections of my article where each brand was discussed in detail.
Final E3 status
Explore – what has worked for others.
Examine – how similar elements are performing in your case.
Extrapolate – the data to find viable conditions for probable success and further testing.
What did I aim to achieve with this Test?
Using Google Analytics, these were the areas that were expected to show marked improvement from the control:
- Page Views,
- Unique Page Views,
- Avg. Time on Page, &
- All Users,
- Organic Traffic &
- Returning Users.
And these are the two popular areas which I chose not to focus on:
- Bounce Rate &
Results of E3 CRO
Increasing visitor engagement even as traffic declined.
This is a comparison of two weeks before and after the baseline was altered:
As you can see, results were almost instantaneous.
Since the article had originally begun to rank sometime in December 2016, a one-year comparison would show misleading (positive) results.
But here’s one that compares the Control between January 01, 2017 and February 4, 2017 (35 days) with the Test between February 5 (the day I made the changes) and March 11, 2017 (35 days):
Consequently, my article shot up to the fourth position in a month. With minor tweaks added later, it has wavered between the first and second places till date – my competitor being Amazon.Com.
This is from March 4, 2018:
And this, from March 9, 2018:
The keyword difficulty is shown as less when I am on top presumably because I’m easier to push down than Amazon!
How do I know these results were achieved by inserting those four lines?
Well, since the Test had nothing else added to it until more than two months later, it was either those four lines or Google’s whims.
Thankfully, though, there’s a more concrete way to ascribe the success to E3.
Take a look at the number of clicks the heatmaps show:
(For 1000 page views between February 27 and March 15, 2017 using Hotjar Basic)
Why were Bounce Rate and Exit left out?
First of all, no expert agrees that Google calculates the Bounce Rate shown in Google Analytics to determine a site’s ranking.
That said, both bounce and exit ought to tell you something about the page’s value. If they are too high (as they are in my case), it is generally assumed that the visitors leave because they do not find the content satisfactory.
However, it is important to understand the purpose for which the visitors are acquired, in the first place.
In this case, I wanted my visitors to take a look at my content and then click one of the affiliate links and leave – because they got the info they had come for.
Obviously, not everyone would click a link and leave. Some would simply leave.
Either way, I did not target my visitors so that they would hang around my site reading stuff. I wanted them to leave – after spending some time reading.
In case of highly targeted visitors, there is no particular reason for the reader of Article A to also be interested in Article B.
Of course, there are ways to make a visitor hang around, but for this particular article, that was not my primary purpose.
I would have considered bounce and exit if the experiment had failed.
Since it hadn’t, and since I usually have very little time after getting client-work done as an SEO Content Writer, I decided not to mend what didn’t seem broken.
What were the ‘minor tweaks’ mentioned earlier?
I had originally inserted a chart that compared not just flushability of the brands but other qualities like how much they track, if they are dusty or not and so on.
This was to present the readers with a more complete picture of each litter for a better informed purchase decision.
The original chart had the column names at the top only.
This was inconvenient.
When you are at the 7th brand looking at the green and red cell blocks, you don’t quite remember if the second cell is ‘dusty’ or ‘septic safe’.
Hotjar videos confirmed that this was so – visitors kept scrolling up and down the chart while looking at it.
I simply inserted the quality name in each cell and that solved the problem:
I also experimented with the copy and kept the pdf version of the chart for download after opting in and finally, not being a big fan of growing subscribers that way (shocking, I know!), just kept it in a no opt-in required state.
The font and colours of the original test lines were also changed several times – but I did not measure if that made a significant difference. What can I say – I’m lazy.
This is how they stand at present:
But these are tweaks any blogger will make from time to time. There’s nothing spectacular (or E3) about them (except for one which I may refer to in a later article) – but they do all add up.
So, does this prove that E3 CRO works?
Well, it means E3 CRO works – to prove anything, further testing is obviously needed. No one in their right mind would claim to prove anything after just one case study.
And even that study wasn’t properly done.
Strictly speaking, at the very least, I needed to remove those four lines and collect fresh data for comparison.
The purpose of this case study was to show a distinct possibility that E3 CRO has potential. That you can safely use it to attempt conversion optimization on your own without any significant expense, not even in terms of time and labour.
I’ll be happy if that purpose has been achieved.
Rest is up to you, my readers. If you should be inspired to use the E3 method, comments regarding your experience would certainly help.
Google Analytics. Hotjar Basic. KWFinder.Com