Screw AB Testing, Build A Better Funnel Instead

Chelsea Baldwin
Chelsea Baldwin

AB testing and the tools that go with it are some of the best gifts the conversion gods have blessed us with.

They tell us what’s significant to users on our landing pages, what isn’t, and help us lift conversions.

Because sometimes, even a lift of a small percentage point can be worth tens of thousands of dollars by the end of the year. (Hallelujah, anyone?)

And as marketers, we love this.

We thank the conversion gods every single day for blessing us with such a great ability to make our businesses better, and the opportunity to throw best guesses out the window.

But sometimes, you don’t need to ask the data’s permission to make a change.

You don’t need to run an AA test followed by another AB test to know if a change is worth making. This is what AB testing statistics tell us.

You just need to man up and go with it.

Sometimes, these things are obvious and you can feel it in your gut that you just need to take a deep breath and make whatever funnel-changing alteration you can’t stop thinking about.

But other times, they’re not so obvious and you don’t have any idea of where to start making a better funnel so you can prevent fall-offs and boost your conversion numbers.

So you keep on A/B testing, even if it is a relative waste of time.

Here’s what I mean:


1) You’ve already tested multiple things on each page of your site.

It will never be possible to test absolutely everything on your website, or all the possibilities of change.

They are literally endless, so it’s best to stick with the obvious ones your appointed conversion expert points out to you.

So if you’re a small business that doesn’t have three million pages to work on and you’ve already tested the factors that will make the most difference?

To continue testing means you’re going to feel like you’re doing the same thing, over and over again, for exactly the same results.

That’s not commitment, people, that’s a waste of time.


You don’t want to waste time, do you? We didn’t think so. – image source


So start looking at your funnel as a whole instead.


2) All the numbers are giving you a headache.

And probably because they’re all too similar to differentiate.

A/B testing is not easy, and the numbers are no joke.

When running a single test, you’ve got to think about significance levels, rate of false positives, the difference in and out of peak seasons and their implications, and test size samples.


It. Just. Hurts. – image source


Even the geekiest of math whizzes will get a headache if they’re desperately trying to dig through this data of test after test, but nothing of great significance is happening.

For this, it’s time to do something a little more drastic.

Set the addiction to the data aside for a week or two and make a significant funnel change just to see what happens.

It might be good, it might not.

But it’ll give you a fresh perspective on where else to start making changes, and on what.


3) You’ve got a gut feeling your high bounce rates have nothing to do with button color.

And that’s the insistence you’ve currently got on chasing small wins via minor changes is making you lose out on the bigger picture win that’s just begging to be unleashed.

After all, even when button color can make a difference, it’s only one small grape vine.

So why pick the grape vine totally dry when you can explore the entire vineyard across the street?

More grapes (and therefore more wine) are always better, right?

For example, maybe you haven’t tested all the colors of the rainbow on the elements you want to stick out on your page. (Who has?)

But you’ve got this gut feeling that instead, you need to change your messaging completely because the info you’re collecting from your customer base is suggesting you’re writing everything for the wrong people.

Stop it with the split testing and just reform your messaging already.

It’ll give you more gumption than even the best color scheme in the world ever could.


4) There’s a huge fall off at one obvious point in your funnel.


Where are they all going? 🙁 – image source


When people are incredibly interested in what you have to offer, but fall out of your funnel in droves all at the same point, it’s glaringly obvious that something’s gotta give.

And it’s probably not the placement of your CTA form or how many required information fields you have.

(Unless you’ve got 20 required fields, then you’ll want to test that against five or less.)

Instead of A/B testing small elements that can be changed and left to chance at that particular point in your funnel, there needs to be a core, fundamental change with how you engage with your prospects at that point.

For example, you might have something that’s totally turning them off, so you need to survey them to figure out what that is and remove it.

Or some part of your offering might be confusing, so you either need to provide more education or explain it in simpler terms.

Or maybe you’ll need to hire more customer support to proactively get in touch with your prospects at that point so they can do more hand-holding so the droves that are suddenly falling off turn into droves of new customers instead.

Do a survey or commit yourselves to proactively being in touch with prospects before.

And after they reach that point in the funnel to figure out how you can optimize your brand experience at that point to stop the drop offs.

This is something A/B testing simply can not tell you.


5) You’ve got low traffic and have to wait a long, long time for a test to show results.

Peep Laja of ConversionXL suggests that you should have 250 full-fledged conversions from a test before you ever sit down to judge the results of it.

That’s 250 brand new customers… not 250 visitors, not 250 CTA clicks, but 250 new people buying in regardless of how long it takes.


> Wow. – image source


If you’ve got low traffic, you’ll never get the results you need in time to make the impact on your business that you’re looking for.

And let me be the first to say, there is absolutely no shame in having low traffic.

Everybody starts there.

But also, many business models actually don’t need high traffic to be successful.

With my own writing and consulting business, for example, I don’t even want to know how many years it would take for me to cycle through 250 clients wanting to pay me for my services.

I might not even go through that many in my entire career. Sheesh.

But, you do need that traffic for A/B testing to be successful.

So if you’ve got a site with low traffic simply because you don’t need a site with boatloads of it, go ahead and put your energy towards optimizing your funnel rather than worrying about percentage-point conversion optimization.


6) You’ve got a million and a half changes you want to implement before the year is over.

Let’s say you know certain pages (or your whole site) is a little sub-par and you know you need a re-vamp.

You’ve listed out the things you want to change, and while it might not quite be an entire website re-design (it very well might be), it’s more things than you’ve got the time to test and get results from.

Please: do not waste your time asking the data’s permission in this instance.

Just go with the changes.

Because if you wait on the data to tell you what to do?

With the speed that the internet changes, you’ll be needing another full-out redesign before the tests on your current ideas are over.

Because yes, even though they might be basic shifts on your landing pages, your website pages are still very much a part of your funnel.

So, particularly if the changes have to do with optimizing your brand experience (rather that conversions only), those are changes that you need to make to strengthen your brand and increase loyalty over time.



Let’s take a look at Plaxo’s redesign.


For example, what if in this instance instead of going for the all-out redesign, Plaxo had insisted on A/B testing everything first?

They’d have to test the 1) messaging of their main headline, 2) shortening their tagline, 3) showing product-based images rather than client logos, 4) placing a CTA above the fold, 5) decluttering their top menu, and more.

And because all these elements fall within the same screen, they’d have to test one at a time until they got them all tested, which would take an eternity.


7) You testing tool can’t call a clear winner.

Which testing tool is it? – image source


This might mean that you’re testing the wrong things on your designed landing page, yes, but there’s a higher chance that it means your drop-offs are coming from more than just not having a readily-available CTA at the exact moment a visitor wants it.

So when your testing data isn’t yielding any kind of significant results, and you just keep going through the motions of constant A/B testing (like a good marketer is supposed to) it might be time to hang up the hat of page-based A/B testing for a while and start concentrating on your funnel as a whole to see what kind of results you can get there.


So… How exactly do you optimize your funnel?

Many marketers know exactly where they need to start optimizing their funnel, but many don’t.

This could be an entire post (or full-out ebook) in and of itself, but here’s some ideas to get you going in the right direction:

1) Do a copywriting audit.

Copywriting looks very different to trained and untrained eyes.

And even though your copywriting might not be totally bad in and of itself, if it isn’t deliberately crafted to evoke emotions and feelings of attachment from your visitors, you’re going to be losing a lot more customers than you should.

To start, hire a high-end copywriter to do a simple audit for you and give you suggestions on what’s missing with your blog posts, landing pages, managed ppc ads, and email campaigns.

They may also be able to give you valuable insights about why people are dropping off that you haven’t thought about before.

2) Identify your funnel’s major drop-out points.

After which emails do the majority of your unsubscribes unsubscribe?

Why do droves of people opt-in for your first ebook, but not the second or third one?

Identify the points in your funnel where people stop converting as highly as you’d like them to, and find out directly from them (via surveys) why they’re not converting.

You might need to provide more value in your first ebook to create a desire for the second one, for example.

Or you might be sending a sales-y email way too soon, before you’ve properly established loyalty and trust.

Find out where people are dropping off, figure out why, and experiment with changes to stop that from happening.

3) Add more education.

This will help you better qualify prospects and increase your brand loyalty.

At a surface level, it can mean writing an ebook, making an explainer video, or putting together a beneficial email series that helps people for free rather than just waiting on your landing ads and landing pages to do all the selling for you.


email marketing

Take a look at Hubspot’s “Featured Download” in the bottom right corner.


Adding education doesn’t have to be hard.

Since I was reading a blog post on email marketing campaigns, for example, HubSpot knows it’s something I’m interested in, so they offer me an educational piece in the bottom right corner to help me with my goals and make me a more qualified lead for their product.


Just Make the Changes, Already

The conversion gods didn’t bless us with A/B testing and A/B testing software to limit us.

They gave them to us to empower us, but only in certain situations.

Think of them more as a sword to use in particular battles rather than some sort of holy knife that you’ve got to use day in and day out for every vegetable you need to cut.

They’re great to use at times, but in other times, using them will only put you behind.

So if they’ve stopped serving you in the way they used to and you still need more conversions than you’re getting?

Start looking at your funnel.

P.S. Did you learn something new? Share it with your followers so you look even smarter :)