Optimization and Split Testing: Losing Money? How to Troubleshoot a Marketing Campaign
The whole point of a campaign is to become profitable.
You’re investing your money into a machine (your campaign). Hopefully, that machine spits out enough money for you to earn a profit.
It sounds simple enough, but most of the time these campaigns don’t start making money immediately.
Creating a profitable campaign is more like having a collection of different parts. And no instruction manual.
These parts include:
- Your offer
- Your landing page
- Your targeting
- Your creatives
- Your angles
- And much, much more.
Your job is to find the right components and combine them together in a way that generates profit.
It’s kind of like putting a puzzle together.
There’s one big difference though. The only thing you need to put a puzzle together is time.
You need money to run campaigns. Money is the fuel to the machine.
I know it’s scary to lose money on a campaign before you’ve found any winners yet.
Think about it like this: You’re not losing money, you’re buying data.
That valuable data is going to give you clues on how to get this campaign profitable.
This is when you go into the troubleshooting stage.
Troubleshooting means that you’re trying to figure out what problems to fix. Split-tests are the experiments you conduct to fix the problem.
Launching a campaign’s a fucking rush.
But you have to troubleshoot a campaign and get it profitable before your money runs out.
The question is, how do you troubleshoot a campaign? What do you do when everything’s messed up and nothing is working?
In this article, I’m going to cover a framework that you can use to troubleshoot a campaign.
A Framework for Troubleshooting
The following is a framework you can use. Keep in mind that the term campaign is a broad one.
And remember that how you optimize a push campaign is going to be different from how you optimize an e-commerce product that you’re dropshipping.
1. Pick the Most Valuable Area of Your Campaign to Focus On
You have fires burning everywhere.
It can feel overwhelming because there are so many things to test, and you’re losing money.
Take a step back. Breathe.
What’s The ONE Thing that you can focus on right now?
What area of your campaign can you focus on right now that will make the biggest impact?
I break campaigns down into broad categories to stay organized.
- Is it the offer?
- Is the angle not hitting?
- Are your creatives performing horribly?
- Is it the landing page?
- Is your targeting wrong?
Does changing the color of the call to action on a landing page make a difference? Sure.
But it doesn’t have as much impact as other variables in the early phases of testing.
There is a hierarchy when it comes to testing or an order. Everyone has their own order.
Here’s the structure that I’ve used for the past 12 years:
1. The Offer – What are you selling?
The offer is always #1. When Apple came out with the original iPhone, they could’ve had the worst advertising campaign in the world. It would’ve still sold out because the offer was that great.
An example could be a pillow for $50. Your average customer knows they can go to the store and buy a pillow for $10. You’re going to market this well for someone to purchase it for 5x more.
2. The Angle – What’s the “hook”? How are you getting people’s attention?
This pillow helps with neck pain. This pillow is adjustable to fit your shape. This pillow is 100% organic and helps the Earth.
There are multiple reasons someone might want to buy this pillow. You have to figure out what’s the #1 pain point.
3. Creatives – The ads you’re using.
For most of you, it’s going to be the images and videos you use on your paid traffic source. The goal at this stage should be to capture their attention, show some quick benefits, and then give them a call to action.
4. The Targeting – Who are you showing this to?
This depends on the traffic platform. On native this can be which placements or siteID you’re targeting.
On Facebook, you could create different ad sets based on interests: biohackers vs. office workers vs. new mothers etc.
You need to narrow down your audience to the most profitable segment as fast as possible for your testing.
Lets say you’re doing native ads and you’re promoting a men haircutting device. It would make sense to narrow the targeting first to male-oriented websites like ESPN and GQ, and figure out the best ads / landing pages there.
Once you’ve optimized your creatives, then you can expand your targeting to more gender-neutral sites.
5. Landing Page / Pre-Sell Page – The website that’s doing the selling.
Just because I listed it last does not mean that it’s not important.
It’s because your angle / creatives / targeting has to be on point for people to even see your landing page.
Pro Tip: I’ve found that testing countries makes a huge impact due to differences in processing, fulfillment, and the cultural impact of angles.
You’re not always going to use the same testing methodology for every campaign because they’re all different.
- Sometimes the offer is unique – There might only be one offer for a COVID-19 Disinfectant lead generation so you can skip the offer testing phase.
- The data can tell you immediately which area is horrible – The offer’s converting alright. The ads are doing alright.
- Holy shit I’m getting an awful add to cart percent on my Shopify landing page. Lets focus on that.
- Some variables are “proven.” – Sometimes I’ll skip testing a variable in the early stages because it’s “proven.” I’ll test a landing page that I see all over the spy tools. And I know people are doing huge volume on the offer using this landing page.
- I can laser-focus my resources on finding the perfect offer/ads. I’ll test the landing page later on in future phases.
Anyways, I know some of you want an exact answer.
“Always test the offer first”
“Always test the ads first”
But campaign troubleshooting is both a science and an art.
2. Troubleshoot the Area You Picked
OK, so you’ve decided on an area to focus on.
Some variables are easier than others to test.
For example, you want to test your offer.
It can be as simple as grabbing recommended offers from your affiliate manager and running an even split test with them in your tracker.
Whichever one makes the most money after you gather statistical data, wins.
However, deciding with data is not fun to talk about.
The other week I was troubleshooting a new landing page so I’m going to focus on that area for this article.
You need to realize that the landing page is a large component. To split test a landing page can feel overwhelming when you’re first starting out.
It is helpful to break a landing page down into sub-components such as the graphic below.
It’s better to focus on those specific areas.
You can use this idea with any part of your campaign.
If your Video ads aren’t doing too great, try breaking it down into its components.
Do you see how much more systematic and thorough you can be if you break everything down?
A. The first thing you should do is to look at the traffic data.
You spent money on this data.
You need to look at the numbers and try to understand what story it’s trying to tell you.
Let’s say you’re looking at your landing page, and you’ve noticed the landing page click-through rate is low, or there’s a low percentage for add to cart.
People are interested because they clicked your ad, but something about the landing page made them run away.
So now let us think of possible reasons why.
- There might be an error on the page itself. Maybe your call to action (CTA) buttons are not showing correctly or linking correctly.
- Your ads could be misleading. Your ads say “Hot local moms near you” but the landing page features hot local dads.
- Something’s scaring them off. You’re selling a pillow. The first thing they see on your landing page is the $199 price tag and they’re “NOPING” their way out of there.
Does this make sense?
Look at the data and think of what the problems could be.
B. Are there external sources of information?
The click data isn’t the only source of information about a landing page.
- You can test the landing page speed using GTMetrix.
- You can analyze the visitors on my page using Google Analytics.
- You can install a heatmap on your site to look at user behavior:
- Or maybe you want to hear actual feedback. Go to UserBrain and you can hear what people have to say about your site.
C. Analyze Other Websites for Ideas
You can always go to a competitor’s Facebook page and see what ads they’re running. From there you can see what they’re doing on their pages.
What are they doing that you’re not?
And don’t be myopic and focus only on competitors in your vertical. Instead, look around.
I’m not in the supplement space or the mattress space. However, I always love watching those spaces because I know some of the best direct response marketers operate in those spaces.
I can easily take an idea from there and bring it into my own campaigns.
This one site is using quizzes to figure out the best product. Maybe I could use that technique in my own campaign.
D. Finalize the Split Tests
By now you should have a ton of ideas on what you want to split test.
It’s time for you to narrow it down. Remember that you’re limited by time and budget. If you’re on a smaller budget, then you can afford to do a few tests due to statistical significance.
Make it count.
E. Execute the Split Tests
Some tests are simple.
If I want to split test headlines on my landing page, I can have that test up and running in under ten minutes.
However, some tests are beyond my capability.
What if I discovered that my website looks fucked up on mobile.
I’m not a programmer and don’t have the ability to change certain things. Even if I did, it might not be worth my time compared to delegating it to someone else.
One pro tip I can give you on delegating is to use Loom.com to record a quick screencast and explain to them what I want.
However, when you delegate, there’s going to be a waiting period. You’ll have to wait for the person to execute the test, do some quality control, and then wait some more for the data to come in.
During this time period, you can work in another area of your campaign.
F. Analyze Results
It has been a few days and now your data should be in.
Did the test work or not?
At some point you’re going to have a make a decision: do you keep testing, or do you quit and try something else?
I don’t have an exact formula, but there are some questions that I ask myself:
- Are my split tests making a difference and moving the needle?
- Are other people making it work? If they are, how many tests do I have left?
- What is the potential of this campaign if it hits?
- What variables are there left to test? For example, if you’re down 50% ROI and you haven’t tested a ton – that’s amazing. It tells me that the campaign at least converts and the only way to go is up. However I’ve applied most of my optimization strategies and there’s not much left I can do, then it’s probably time to move on.
- Setting a budget beforehand – This is a concept that I learned from playing Poker. I always enter a campaign with a “Stop-Loss.” It’s easy to get emotional and tied to a campaign that you’ve spent so much energy on. Come up with a rule beforehand.
If this campaign is not breaking even after $2,000 spend, I’m finding something else.
The Testing Never Stops
I hate the concept of passive income.
Does it exist? Sure.
If you own a ton of dividend stocks and you get a nice payment at the end of each quarter.
Or if you own some property and have a kickass property manager handle everything.
But when it comes to running an affiliate or e-commerce campaigns, I don’t believe that it’s ever going to be passive.
Once you start making money, the sharks are coming out. The more you scale, the more your business is going to be exposed to competitors.
The more you scale, the more banner blindness kicks in.
The testing never stops.
The world is constantly changing. You’re making $1,500 a day and then the world gets shut down because of the COVID-19 virus. How are you going to adapt to it?
The competitors are innovating. Are you going to run the same way when others are jumping ahead?
It’s frustrating to lose money. Trust me, I get it.
You have to start training your mind to think differently. If a campaign’s not profitable, that’s a puzzle that you can solve. If you solve it, then you’ve gained some knowledge that can be applied in the future.
The key to becoming better at troubleshooting campaigns is experience.
The key to the experience is taking action and failing.