New Shopify Certification now available: Liquid Storefronts for Theme Developers

I spent over 10 Mn USD on Facebook Ads - Here's what I've learnt

Ravish_Monk
Visitor
3 0 6

So I've been getting a ton of DMs regarding FB Ads from members of different communities & instead of replying to all of then 1v1, I'm writing this post as a quick overview of best practices based on my personal experience.

Background:
I've lead a digital marketing team at a B2C startup for about 4 years now with a core focus on optimizing advertising ROI. We've scaled 7x in advertising spends while also improving ROI (ROAS) steadily with the help of product, tech & design changes. Now I'm building an upsell & cross sell app for Shopify stores called Monk (link).

Disclaimer:
FB Ads is not a silver bullet.

An FB ads campaign can be imagined as a set of levers. Our aim is to identify the top 5 levers that move our north star metric - ROAS. 80% of impact will be generated by iterating and focusing only on these 5.
Possible levers are - Campaign objective, Optimization event, Creative, Creative type, Audience targeting, Age, Gender, Location, Landing Page, Platform, Placement, Device , OS, CBO etc

In the 4 years, my team and I have meticulously tested each and every lever & here are my suggestion on what you should be focusing on the most:

1. Know your Target Audience
Who is most likely to buy your product?
What does their demographic look like?
You need to thoroughly know who's buying what you're selling - It will help you fix 4 big parameters in your campaigns: device type, age group, gender & location.
Eg: We noticed that our highest converting cohort was women, age≥30 living in metropolitan cities. This lead to us creating separate campaigns specifically for this TG.

2. Creative, Creative, Creative
The most underrated lever of all. I cannot stress this enough, you need to invest sizable time and effort on your creative strategy. The right creative theme can do wonders for the growth of your business. I've had times where our ROAS shot up 4x the day we added a new creative. No other change was made.

2a. If you use a catalog
The images you add in your catalog need to be of high quality. Avoid images that are blurry, oddly lit or are taken from weird angles, I suggest you should invest in a professional studio to take these images. Don't let it seem amateur (especially if you're drop shipping), it's the last place you want to cut costs.
2b. If you use individual creatives
Video >> Carousel > Single Image.
Video is KING.
The easiest way to create a video is to choose your images, turn them into a slideshow & add simple text overlay to convey a discount offer or a key product feature. Tools like veed.io can help do the job. Carousels are also a good choice provided you have high quality product images & several variants of the same product.

Another key aspect is that in certain geographies & TGs - adding human images to your creatives works wonders. I've run several A/B tests on this singular aspect & the results are undenaiable. Instead of an image of a t shirt, show an image of a human wearing that t-shirt. Instead of just showing a piece of jewellery, show an image of a human wearing it. I understand that all of us don't have the means to get a model and book a studio and get a photoshoot done, here I recommend using stock footage + some shrewd editing skills on Canva to get the job done.
Eg. If you cannot show a woman wearing the necklace you're selling, edit your product image to just add the face or neck of a woman on the left hand side with the necklace on the right, highly likely this will get lower CPCs than just an image of the necklace, add transition effects like a fade in fade out & text to it and you have you 10 sec video ad up and running!

3. Campaign Optimization
The campaign objective & the event you optimize in your adset for will significantly impact your conversion rates & hence ROAS. Unless you're an absolute beginner with 0 orders, I recommend always choosing Conversions as your campaign objective. In terms of which event to optimize for - look for events that have at least 10-20 events per day for the algorithm to learn. So start off with optimizing adset for the store page view event. Once your next step in the funnel - product page view is >10/day, change it to product page view & once you have >10/day add to carts change it to add to cart and so on for checkout click and then final purchase.

4. Audience Targeting
Lookalikes work. If done right.
For lookalikes you need a seed custom audience of minimum 1500 people/events. These should be 1500 people who've shown a clear purchase intent (like add to cart or checkout click). Till you don't have this set of 1500 people, I recommend trying interest based audiences to see what fits for your business and TG. But once you do cross this threshold of 1500 add to carts, instantly move away from Interest Based targeting to Lookalikes.

5. Store Performance
Once the customer is on your store, it's all about two things - store performance & conversion rate optimization.
If your theme is too heavy & you've added a lot of custom code to your store you'll see a significant drop in clicks to landing page views due to poor page load speeds. I personally rely on Google Page Speed Insights to monitor this. A high performing store that loads quickly is extremely important to reduce drop offs & improve ROAS (especially in developing markets with patchy internet) - the key metric you're looking at here is Time to Interactive. Anything sub 2.5 seconds is exceptional but above 5 seconds would need some optimization.
Club this with working on improving the conversion rate/AOV of your store - customer reviews, Instagram feed, video testimonials, sales pop up, upsell & cross sells (Monk) etc - all of which that can be done via several free apps on the Shopify App store. Your ROAS will simply not move if you don't focus on optimizing your store, there's only so much you can do from FB optimization.

If you found this helpful - do you recommend I write detailed individual posts on each of these levers? Or maybe even think of doing a webinar?

 
Replies 0 (0)