Five Tactics to Profitably Scale Facebook Ads for Your Shopify Store

Wondering how to acquire more customers within margin for your Shopify store? Tired of shooting in the dark on your ad tests? Don’t worry! I’ve compiled my 5 favorite Facebook ad tests for Shopify stores. Read this guide, and take the guesswork out of your next five tests.

Five-Tactics-to-Profitably-Scale-Facebook-Ads-for-Your-Shopify-Store.jpg

#1 – Test More Lookalike Audiences

When running Facebook ads for ecommerce, most people only test a handful of lookalike audiences. Usually, people run tests based on buyers or site visitors. 

What I’ve found through my testing methods is that interest based traffic is terrible 90% of the time and typically makes up the worst portion of your traffic portfolio. Here’s my theory as to why…

Let’s say you’re running a campaign targeted to people interested in dogs and cats. You’re only telling Facebook “show my ad to people interested in this subject”.

When you run a campaign targeted towards a lookalike of people who’ve bought dog food, or signed up for a dog training newsletter, you’re telling Facebook “show my ad to people with similar characteristics to my buyers or my leads”.

In short, interests don’t necessarily include a propensity to take action as a characteristic, lookalikes do.

In light of this, I recommend testing as many different lookalike audiences as you can, as long as the seed audience is larger than 1,000.

Here’s a cheat sheet of potential lookalikes to get you started:

  1. All buyers
  2. All leads
  3. All leads + buyers
  4. Buyers last 30 days
  5. Buyers last 60 days
  6. Buyers last 90 days
  7. Site visitors 7 days
  8. Site visitors 14 days
  9. Site visitors 30 days
  10. Site visitors 60 days
  11. High value buyers export from Shopify (When exporting a CSV of customers from Shopify, you’ll need to filter by “total spent”, and look at your historic data to see what the high end of your average cart value is)
  12. High value buyers via Facebook custom audience
  13. High volume buyers export from Shopify (users who’ve bought more than once)
  14. High volume buyers via custom audience
  15. Any user who saved one of your ads (when making a custom audience, click “engagement on Facebook”, then “page”, then include “people who saved your page or any post”)
  16. 95% video viewers
  17. 75% video viewers
  18. Mobile buyer custom audiencemobilebuyercustomaudience.png
  19. Desktop buyer custom audience
    desktopbuyercustomaudience.png

#2 – Lookalike “Stacking” + Age and Gender Segmentation

As Facebook’s oCPM algorithm gets more and more refined, the recommended audience sizes for cold traffic go up and up.

When I first started advertising on Facebook, we would typically run audience sizes of 250k-500k. Then it became 1-3 million. Now I’m having success with 3-6 million. There is typically an inverse correlation on Facebook between audience size and cost.

A way I’ve found to keep audience sizes high while also mitigating the risk of under-segmenting is with “lookalike stacking”.

Take your lookalikes from step 1. Include 5-10 of these audiences in one adset. Next, you’ll need to determine the best demographic segmentations to use.

The old way of finding your correct age, gender, placement, and device segmentations involved exporting a CSV into excel and running a pivot table. For most people, this is a confusing, counter-intuitive and time consuming process.

You know how long it took me to figure out how to make my first pivot table? Months! And once I got used to doing them, it still took hours to get all of the data together to a point where it made sense.

Luckily there’s a new and easy way to find the correct segmentations for your adsets. If you log in to your FunnelDash account and go to the Facebook Ads Audit Dashboard you can easily find the most efficient and most profitable, age, gender, and placement segmentations to use for your new campaign.

Then go back to your adset and start segmenting. For each stack of lookalikes I’d recommend 3-4 copies of your adsets, with different age and gender segmentations.

#3 – Conversion Event Test For Cold Traffic

In my experience, Ecommerce advertisers typically optimize on purchase event. What they don’t know, or haven’t tested, is that results actually vary depending on what conversion event you optimize for with the conversion objective. Each conversion event gives you a slightly different subset of your audience and therefore gives you a different cost.

For cold traffic I recommend running a “conversion event” test and testing:

  • Purchases
  • Add to cart
  • Custom conversion for a specific product buy
  • Custom conversion for a specific product add to cart

You can make these custom conversions by using a product ID & a specific event.

To set up these events, first install the Flexify plugin for Shopify. Then, go to your Shopify store with chrome and pixel helper installed, go to the specific product page, click pixel helper, and you should see a content ID.

Back in Facebook when you create your custom conversion, choose event, and then in the box below “Add To Cart” or “Purchase”, drop in the content ID of your product.

This is a test you’ll run once, for a week on your best audience. Moving forward I’d recommend only running the winning conversion event.

#4 – Multi-variate Retargeting Strategy (Bid Type Test + Audience Test)

This is a strategy I’ve tested across multiple accounts and it always seems to work.

It runs off a similar theory to the conversion event test. If you have a retargeting audience of site visitors, and you have

  • one campaign that bids for conversions
  • one that bids for clicks (known as “traffic” as in ads manager”
  • one for video views
  • one for post engagements

If you’re targeting the same site visitors audience for each of these campaigns, you’re going to reach a slightly different subset of that audience in each, and therefore, have a different CPA, ROI & ROAS for each as well.

To take full advantage of this strategy you want to do a retargeting audience test at the same time you’re doing a bid type test.

Here are some audiences I’ve tested with this strategy:

  1. 7 days site visitors
  2. 14 days site visitors excluding 7
  3. 30 days excluding 14
  4. All Leads
  5. Users who’ve bought product A but not product B
  6. 75% site visitors
  7. 95% site visitors
  8. Page likes
  9. Anyone who saved a post

For each audience you test you’ll want to test:

  1. Bidding for purchases (web conversions objective optimizing for purchase event)
    optimizing_for_purchases.png
  2. Bidding for add to cart (web conversions campaign objective optimizing for add to cart event)
  3. Bidding for clicks (traffic campaign objective)
    clicks=traffic.png
  4. Bidding for video views
    videoviews.png
  5. Bidding for page engagements
    postengagementscreenshot.png
  6. Bidding for reach
    bidding_for_reach.png
  7. Bidding for impressions
    impressionsbid.png

I know this could easily be 50+ adsets but trust me, it’s worth it. Launch each adset at a minimal budget (however much you want to get a sale for) and let them run for a day or two. Each audience should have a winning bid type (winning adset) and I think you’ll be surprised at the ROI of the winning adsets.

#5 – Increase Your Reach On Winning Retargeting Adsets With Manual Bids

Have you ever watched Mad Men? Did you know Don Draper is actually based off of a real 1950s ad man, Rosser Reeves?

This next ad tactic comes from a bit of his advice in the book “Reality in Advertising”. (By the way, I recommend his book, and three others as required reading in this blog post.)

In the book, Rosser talks about dispersion in media buying, and cites an example where his returns went through the roof when he began to buy up a higher quantity of cheap radio spots and increased the dispersion (reach) of his message instead of spending money to increase frequency. Here’s the quote from Rosser:

The ideal dispersion, of course, would be to reach 100% of the people and then repeat with as much frequency as the budget will permit. There is no such buy; but until an advertiser can reach maximum audience, 80% to 90%, he must keep reaching out for more and more different people. Then, and only then, should he begin to add frequency”

To translate what Rosser is saying to 2017, dispersion can be seen as impression share, or reach / audience size as a percentage.

To implement this strategy select your winning adsets from step 4.

Edit your ads as needed.

Take note of the projected reach, and then set a manual bid that increases your projected reach by 20%. For example, if Facebook tells you your projected reach is 1,000, set a manual bid that increases that number to 1200.

Do this for every adset and wait a day or two. If your results hold steady, increase your manual bid again. Keep doing this until you reach 80% impression share. Once you reach an impression share of 80% or more, increase your budget until your frequency is 3-5 over a 7 day period.

Wondering which of these tests to tackle first? Just take a look results for the last 30 days and pinpoint possible areas of improvement. Wondering how to expand your audience targeting for prospecting campaigns? Start with test 1.

Is your cost per acquisition on cold traffic too high? Go to test 2. Remarketing need a boost? Start with the bid type test and then once you have winners move to manual bids.

Want a Quick Reference Guide for Running These Tests?

Download our guide: “5 Tests to Run on Your Facebook Ads to Increase Your Ecommerce Profits.”

Click on the button below to get your copy now.

 These techniques are by no means foolproof, but they have gotten me some amazing results in the past. For example, I achieved a 700% ROAS last cyber Monday for one Shopify store owner by employing a combination of bid type testing and audience testing for remarketing.

Remember, Facebook rewards proactive campaign management… So, Happy Testing!

Have you run any of these tests? Are you excited to run one in particular? Tell us about it in the comments.


>