Darius Kunca | Co-founder and Lead Strategist at AdKings Agency
July 06, 2019 - 9 min read
[Case Study] E-Commerce Cosmetics Retailer Reverses Decreasing ROAS Trend, Increases Facebook Ad Spend x6, Revenue from $1,396.25 to $11,084.73 and ROAS by 132% in 60 Days using the 'ICE Growth-hacking' Framework.
✔ How to use the ICE Growth-hacking framework to reverse months of declining Facebook Ads return on ad spend (ROAS).
✔ The ‘Sprint’ method: A simple 2-step strategy that helps you structure your work and execute ideas the most efficient way.
✔ How we have Scaled The Business from $1,396.25 to $11,084.73 a day in as little as 7 days.
Majority of Ecommerce owners in 2019 identified these 3 scary trends…
❌ Skyrocketing Increase in Competition
❌ Diminishing in Marketing ROI
❌ Plummeting Profit Margins
These are three key problems that one of our Facebook Ads clients faced, but we solved. as well and that we have managed to solve.
Last year they have started really well, and in during Q4 our client managed to scale up to $100-150K in sales revenue per month. mark.
But 2019 wasn’t that welcoming…
They were hit with a sudden increase in the market competition by an influx of Chinese brands. Which meant…
They Were Forced to Downscale from $2K/day Ad spend to Just $300/day
Their ROAS started to plummet from 2.0 (3.0 at Shop Level due to Underreporting of Facebook) to only 1.7 (2.5 Shop Level due to Underreporting of Facebook)
And they were barely profitable…
What’s worse - they were forced to downscale from ~$2000 in Ad Spend on Facebook to just $400 per day… This meant they were having a hard times paying-off all the mounting business expenses.
It was clear that things took a sudden wrong turn...
And this is where we came in.
It was our job to fix that and put them on the right track.
Which I would dare to say we did with a huge success.
We managed to turn that $300-500 at 1.72 ROAS (2.5 at Shop Level) to $3k per Day at 2.2-2.3 ROAS (3.3 at Shop Level)
And the only thing stopping us from scaling more is that they do not have enough stock!
That’s a good problem to have, especially when just 2 months prior, they were raising questions whether they will be able to sell out their inventory at all.
How we did it, STEP 1: Analysis
To solve the problem, you need to understand the problem… (Now I’m sounding like Aristotle, but bear with me…)
What we did was take a look at 2018 sales data vs 2019.data. Comparing two years, one then the things were going well and another one – then things went downhill. This meant we were comparing like with like and could spot where things went awry.
Taking a look at these graphs we were able to identify the time period from which the problem originated and started at around May 2019th.
Afterward, the next step was to compare problematic period micro-data vs the one then performance was good. This way getting the clear picture which KPI's have changed. Whatever it decreases in CTR, increase in Link Click Prices or decrease in Conversion Rate that might be causing the problems.
Data From Then Things Were Doing Good During 2018th...
VS During Problematic Period in 2019th...
We noticed that:
- CPM’s have increased
- Link CPC’s have increased
- Ad CTR’s have increased
- Add to Cart, Initiate Checkout and Purchase Prices have increased
What was interesting, that while ads CTR’s had increased, the core metric of how much does it cost to drive the person to the website actually have INCREASED…
This immediately raised these possibilities for us:
- Creative burnout (would be strange to see an increase in CTR’s if that would be the case, but creative haven’t been changed for 3-4 months, so it for sure would be something we would have to test.)
- Wrong Audiences or Audience Burnout
- Lookalike Audiences Collected the Wrong/Old Seed Data
- Frequency problem of hitting just a small part (~1%) of the target audience over and over again
- Algorithm assigning low Quality Scores and Negative Feedback
- Algorithm malfunctions
- Algorithm deeming the website experience low quality
STEP 2: Brainstorming Solutions and Implementing ICE Score Framework
Now for each of possible problem, we had to come up with solutions list. Here’s a sample of what we came up with: which looked just like that:
- Prepare 3-4 new videos, with new sales angles
- Prepare 2 Instagram Stories to test different types of creatives
- Test User Generated Content Videos and Reviews to Cold Audience
Wrong Audiences or Audiences Burnout:
- Test new interests
- Test new countries
- Test new targeting methods
Lookalike Audiences Collected the Wrong/Old Seed Data
- Test smaller time frame windows for LAL’s like 7d, 14d, 30d not just 90-180d
Frequency problem of hitting just a small part (~1%) of the target audience over and over again
- Test aggressive exclusions of 180 day Visitors and 365 Facebook and Instagram Engagers
Algorithm assigning low Quality Scores and Negative Feedback
- Test different products, check for the feedback score
- Test new advertising accounts
- Test new pixel
Algorithm deeming the website experience low quality
- Testing 3rd party landing pages
- Testing custom coded landing pages
Now we have evaluated them using something called ICE score methodology on 3 core elements in score 1-10:
I – Impact (how much difference could it make)
C – Confidence (how likely it’s going to work)
E – Ease (how easy it is to implement)
Then we divided them by the 3 and get ICE score.
This is how we prioritized the things to do in Sprint 1 and how to start tackling the problem in the most effective fashion.
STEP 3: Sprint 1 and Learnings From It
Sprint 1 consists of the highest ICE score activities on that list. Majority of them are also the easiest to implement, meaning that they will take less time to test.
The core focus here was on testing:
- New creatives
- New sales angles
- New audiences
- Smaller time frame LAL’s
- Aggressive exclusions testing
- New Facebook accounts
This sprint took one month to complete. While we didn’t completely manage to reverse the ROAS decrease at this sprint we manage to get a few very important key learnings that helped to improve the final result.
3-key learnings from Sprint 1 were...
- New untapped LAL’s and smaller time frame LAL’s are performing slightly better (about 10% improvement).
- Aggressive exclusions are helping to reach a new audience and are performing slightly better (about 10% improvement).
- New Facebook account seems to help to decrease the CPA to what it was before the problem occurred for about 1-2 weeks, afterward it starts to go up once again. (About 30-40% improvement for 1-2 weeks)
And the last one is what helped us to raise the question…
Why is it that a new account starts well and goes off as time goes?
After deliberate consideration and knowing that the owners did a website redesign 6 months back we have raised a question about it. We haven’t considered it to be a serious problem as conversion rate stayed the same after the website redesign, now we have raised a question about Facebook starting to throttle this business and this was a possible culprit.
This was the key finding that inspired a lot of Sprint 2 actions.
STEP 4: Sprint 2 and Learnings From It
Sprint 2 consisted of updated info from Sprint 1 and other ideas that came after the Sprint 2.
The core ideas tested here were alternative e-Commerce funnels to facilitate higher engagement, thus sending signals to Facebook that people are liking the content and what they see.
Sprint 2 focused on:
- Testing Alternative Products
- Quiz Funnel
- Ecommerce Custom Programmed Landing Page (maximized for speed)
- Advertorial Funnel
As you can see these activities are much harder and more costly to implement and we proceeded to them only after exhausting the easiest and most frequent solutions to similar problems.
We started to test the things from the Quiz funnel and noticed x2 times higher Link CTR’s. There also was a huge decrease in Link Click prices and showed us that we are moving into the right direction! But sadly, the final conversion rate hasn't been up to par to where we would like to see it.
STEP 5: Breakthrough Point and What Have Enabled It?..
Over the Sprint 2, we have experienced the breakthrough moment…
After testing the quiz the next step was testing Custom Programmed, Speed and Conversion Rate Optimized Landing Page.
This is where the magic just happened. After the late-night launch, we have woken up to see amazing 2.3 ROAS!
The conversion rate went up from already amazing 4-4.5% to 8%... For the cold audiences…
Imagine what YOU could do with conversion rates like that?
What is interesting we also have seen link CPC prices decrease (matches with our theory of low experience quality on the website).
We have waited for 48h to make sure it’s not just a random fluke and reported our results to the owners.
You bet they were happy! They have never seen conversion rates like that!
STEP 6: Scaling Up Using Our CBO x73 System Approach
And now as we have had a stable footing to kick-off, the fun starts…
Let’s commence rapid scaling – the thing we proud ourselves to be really good at!
First - we have launched a few regular Campaign Budget Optimization (CBO) Campaigns with multiple audiences in them. Some of the targeting Worldwide and some of the hitting BIG5 English speaking countries.
This is what we call stabilization campaigns that give a footing and regular testing for anything that we build on top of it.
The goal of them is to consistently spit out the best audiences for that time frame window and enable us to identify them the quickest.
Afterward, we took each of the well-performing audiences and moved them into our structured CBO x73 System Approach (you can see the diagram on top). Where we are splitting the same audience into 7-to-11 audiences in the same CBO with multiple HIGH manual bids hitting EACH and EVERY customer that has a VERY HIGH chance of a conversion.
With these methods combined, we have managed to quickly hit ~$3000-3500 Daily Ad Spend With ROAS 2.2-2.3 (3.3+ At the Shop Level Due to Underreporting)! That's an improvement from $400-500 a Day at 1.7 ROAS...
That’s one of the more advanced Facebook tricks that I won’t be going in-depth here about. If you are interested to learn more, You Can Learn More in Our Free Facebook Group – click here > https://www.facebook.com/groups/adkingsmastermind/
STEP 7: One Happy Client!
Short things short, after this moment we have blown it out of the box and just within 1 week of scaling, we have reached 10x the scale with around 35% improvement in ROAS!
If you have enjoyed this content, I highly suggest joining our FREE Facebook Group where we share content like this weekly -> https://www.facebook.com/groups/adkingsmastermind/
Want to Know If We Could Be a Good Match to Growth Your eCommerce Store FAST?
Secure your FREE Facebook Ads and "Everest" Growth Framework Call with one of our co-founders - Justin where he will personally take a look at your business, it's problems, opportunities and give you valuable insight and best solutions for your growth at no cost to you.