6 Tips to Help You Improve Your Facebook ROI by Over 300%

by Andrew Krebs-Smith | June 14

facebook ads

Like many companies, we absolutely love Facebook. Why? Because it works! Our clients are showing absolutely enormous ROI on their campaigns. But one thing is incredibly important to note – none of the campaigns started that way.

facebook roiFacebook is set up to get as much money from each advertiser as possible. That means showing your ad to as many people as possible. But you as the advertiser don’t want to pay to show your ad to as many people as possible. You only want to show it to the people who are most likely to buy!

Our job is to scale ROI for our clients. So we thought we’d come up with a list of things you can implement today that will start dramatically improving your ad performance on Facebook. These are the exact same methods that we use to increase our client performance by over 300% on average.

1. Test Your Ads

If you aren’t testing your ads, you are paying too much. Why do you know that your logo makes people click on your ads more than a kitten smoking a cigar? You don’t! (Unless you’ve tested it!) Usually the challenge for our clients at this stage is to widen the breadth of things they test. Don’t test your logo in one ad, and then your logo in grayscale in another ad, and then your logo with a different font in a third ad. You want to widen your “net” as much as possible, to make sure you aren’t missing out on opportunities. Challenge your assumptions!

Want some specific tips? Here are some insights from our ad testing across multiple client campaigns:

  • Pictures of cats sell almost everything. If you can get a picture of a kitten using your product even better. They aren’t usually the top performing picture, but they are ALWAYS in the top 50th percentile of image tests.
  • Your logo never sells anything. Ever. We have never had a client logo win an image test. Why? Because nobody (1) has ever seen your logo so it means nothing to them, and (2) nobody cares. People care about stuff they can relate to. Logos are important, but they aren’t important for garnering the attention of people.
  • Young audiences love this headline: “Every time you get a question wrong on the SATs, a kitten dies. Please, think of all the kittens” – this was for our SAT prep client. Think it doesn’t apply to your brand? We also used it for neckties: “Every time you wear the same tie twice in one month, a kitten dies. Please, think of all the kittens.” Cool right? You are probably thinking – “ok, but I can’t sell my B2B product/service with it.” But how do you know if you haven’t tried? We’ve used that headline to sell ski tickets, business cards, and more. Sometimes it doesn’t work, but we know that only because we tested it.
2. Test Your Audiences

This one might strike you as obvious. Of course you want to test different audiences! But the important part is how you test them. First, make sure you’ve already done image and headline tests. That way, you’ve lowered your CPC so that you can test audiences more effectively. What do I mean by “more effectively?” I mean that to know whether an audience is good or bad, it doesn’t matter how expensive the clicks are, or how big the audience is, or anything else. The only thing that is important when testing audiences is sales data (or conversion data if you aren’t selling things on your site). You need to know which audiences PURCHASE, not which audiences click. Thus, if you want to measure who actually gets all the way through that process, you need to have a pretty large amount of clicks on your ad to get people to check out. You need to achieve some degree of statistical significance, and if you are only getting 5% conversion on your site, that means you need at least 20 clicks just to get one checkout/conversion. And one checkout doesn’t come close to giving you enough data to know if that audience represents a good opportunity. And once you do get enough data to evaluate an opportunity? Then you need to test additional audiences!

Suffice to say, testing audiences is important. But if you want actionable data, you need to do your testing in a structured and sophisticated way so that you don’t (1) get misleading data, or (2) blow your budget.

3. Implement Conversion Tracking

Ok I’m not going to spend a ton of time on this one. If you aren’t doing this already, shame on you! But if that’s the case, you have a huge opportunity. You have the opportunity to stop tracking how many people click on your ads (which means very, very, very, very, very little), and start tracking who is actually buying your stuff or filling out your form. There is plenty of information on the web regarding the why and how of conversion tracking. Once you’ve read that, you can come back and play with the big kids.

4. Implement Site Retargeting

If you aren’t doing this already, this falls into the category of “low-hanging fruit.” It’s also something you should implement ASAP, since it only works moving forward, not retroactively.

Site retargeting basically tracks who has visited your site, and keeps track of them as an audience that you can then target and advertise-to. This is great, because if they didn’t sign up for your newsletter, and didn’t buy your product, how else were you supposed to reach them? Now you can. And although sometimes you can reach these people through email, you have no control over who opens your emails. You can guarantee that people who are on your retargeting list see your ads, so you ensure 100% saturation of your messaging.

5. Implement Lookalike Audiences

This is by far the coolest thing on this list. Without getting too deep into the nitty-gritty, lookalike audiences is basically a way for Facebook to give you people that are very similar to your best customers. The best part? It works really, really well.

How it works: You take your email list of your best customers, and upload it to Facebook. Facebook matches those email addresses with users in it’s database, and then finds users that have very similar traits. You can still use all of the normal targeting parameters, i.e. behaviors, interests, and demographics, but you are starting from a group that is already extremely targeted. Huge win!

6. Test For the Ideal Number of Steps in Your Funnel

Did you know that sometimes actually adding a step into your customer conversion funnel can DECREASE your overall costs? We didn’t either until we started testing it. For certain clients, we actually get much better results when we run ads to an “intermediate lander.” On the other hand, sometimes taking a step OUT of the funnel decreases cost. How are you supposed to know which is better? Test it!

Hint: This strategy usually works best when you can get really cheap clicks. We use intermediate landers most effectively when we are promoting a piece of content for a client. That way we can get cheap clicks to promote the content (via Outbrain, Taboola, or something similar), and if the content is well-written (that’s the important part) it will convert readers to your “real” lander.

Are you excited yet?!?! If you’ve been paying attention, you should be! These strategies can save you tons of money – easily 60-80% of your advertising costs!

Andrew Krebs-Smith
About the author

Andrew Krebs-Smith

Helping B2C Retail/Ecomm companies test, measure, and scale digital marketing. Trying to fix the agency model so that agencies are accountable to every client media dollar they spend.