2018 In Review – Key Learnings on Facebook Ads & Marketing in General

In case you missed the beat, 2018 was a year when online advertising got ever more consolidated to Facebook and Google.

In 2018, Instagram Stories exceeded Snapchat in DAU. Facebook had the toughest year ever. (But nonetheless remains the foremost destination for marketers looking to reach a wide audience at arguably low cost)

For me, 2018 was the most eventful year ever. A year ago, I was freelancing for several startups while growing this blog. 12 months forward, my focus has shifted 110% to working with Taxify’s marketing team. (We’ve grown from 10 people to 30 in just one year, and we’re hiring 🤭) From attending for the first time Cannes Lions festival to last-minute panic-troubleshooting a large-scale marketing operation with Taxify’s French team, 2018 presented plenty of opportunities for learning and growth.

This said…

I took a look back at all the marketing projects I contributed to in 2018 and realised there were some interesting key learnings, potentially useful for other marketers.

Heads-up: a lot of what I thought to be efficient in marketing turned out to work the opposite way. And there’s a surprising amount of obvious best practices that no marketing blogs write about. Go figure…

So, without further ado, here are some of my key insights from 2018 that might be useful to you as well:

Facebook ads get over-optimized all the time. You’ll get better results when you let them run in peace.
Run more-more-more A/B tests about every aspect of your marketing activities. 90% of them fail, 10% will improve your results signifcantly.
Measure everything you do. Yes, even things you think will obviously work. Because sometimes, they don’t.
Facebook video ads don’t work (a shocker!). It’s a bit of an exaggeration, but not by much. Read point 4 of the article for further explanation.
If you spend your time on marketing campaigns, also spend your 💸 on promoting them. + Focus on one big campaign rather than ten small ones.
Create all marketing campaigns, ads, emails, web assets, etc. first for mobile. Get over your desktop-skewed views.

If this list just sounded a bit click-bait-ish, you will find a more in-depth explanation on each of the learnings below.

+ I also did some digging in my notebooks and bookmarks, resulting in a list of the nice stuff I’ve been listening, watching, and reading in 2018. Scroll to the very end of the article for that.

1. Leave ad optimisation to the algorithms

If you’re using online ad channels to promote your product, you likely spend most of your 💸 with Facebook and Google.

Both Facebook and Google have increased their ad placement inventory significantly in 2018. Facebook recently added Facebook Stories ads on top of its existing placements while Google is expanding its Youtube video ad inventory.

Just take a look at all the Facebook ad placement options available in December 2018. A year ago, there were so much fewer ad placements to choose from…

This is getting overwhelming

On one hand, advertising on Facebook and Google is getting increasingly difficult.

You need quite a specialised marketing experience to play around with all the targeting, optimization, and placement options of main online ad platforms.

But there’s also good news – 2018 also brought about a wider use of AI and Machine Learning.

So…

Should you build an AI bot to do the Facebook advertising for you? – No.

Should you use third-party online tools that claim to optimise your ad campaigns for you. – No.

Just let Google and Facebook algorithms do their magic.

One of the articles I wrote in 2018 and like most is the one explaining that you don’t need to over-optimise your Facebook campaigns. Just let Facebook’s algorithms do it for you.

What happens if you optimise your Facebook ad campaign’s delivery schedule, placements, narrow down the targeting to 10k people or less?

I’ve mostly seen this happening:

Over-optimisation will start limiting your results

So how can you set up Google and Facebook ad campaigns that will reach a wide audience interested in your product?

Try these best practices:

Spend at least $20/daily per every new ad set – algorithms need a lot of conversions to optimise your campaigns.
Optimise ad delivery on the right events – if you’re working with low budgets, optimise on landing page views. If you get at least 10 down-the funnel conversions daily per ad set, optimise on signups, etc.
Set up a single ad campaign with a single ad set and 1-3 ads while resisting the urge to fragment your campaigns into scattered low-budget ad sets. This applies to low-budget Facebook ad campaigns.
Don’t A/B test too many elements at once – you won’t have enough results to conclude the test. Here’s my take on Facebook ad A/B testing.
Don’t worry too much about the ad placements, custom delivery schedule, and manual bidding. Use Facebook’s default settings for all these things.
Set up Facebook Pixel to track your campaign’s results and optimise your ad campaigns on conversions. (So that algorithms can learn which type of audience converts on your offer and deliver it to more similar people)
Don’t make changes to your online ad campaigns too often. Adjust your campaigns 24h after setting them up (in case they’re not delivering yet), then let them run for 7 days minimum before making any further edits.

So, to wrap this up: don’t panic when your Facebook ad campaigns are not delivering amazing results 2h after publishing the campaign. Let them run for a while so that algorithms can start working their magic.

2. Run more A/B tests – 90% will fail, 10% will be game-changers

As someone lately explained to me, having 10 x 0.7% incremental improvements in your ad campaigns’ conversion rates results in 2x total improvement.

If you want to improve your marketing results, keep A/B testing new approaches.

However, be prepared to get demotivated once your tests show no significant improvement.

I ran 50+ A/B tests in 2018 and most of them showed either no improvement or negative results.

Quite disappointing, huh?

But that’s just fine.

The remaining 10% of A/B tests brought learnings that improved our marketing results big-time.

There’s an almost infinite pool of things to A/B test. And if you’re working with considerably low marketing budgets, you’re limited to a small number of tests.

So. What should you and what shouldn’t you test?

In my experience, the biggest improvements will come from optimising your online ad visuals. That kind of makes sense – your ad image is the first thing to catch people’s attention. (Also, the first thing that can cause ad fatigue is when people see the same visual over and over again for 10+ times.)

So, start by testing your ad visuals.

At Taxify, testing new ad visuals goes hand-in-hand with updating our branding.

How our carousel ads evolved in 2018

This nonetheless means that you should ONLY test the visual side.

Think about your online ad campaigns as a collection of many different aspects, each contributing to the conversion – your ad image and copy, but also the landing page and later user experience of your product… And the emails you send out. Be on the lookout for ways to improve each step of the conversion funnel.

And don’t just assume that all the best practices shared across marketing blogs work in your case! We’ve discovered many best practices to be either useless or damaging to campaign performance.

3. Measure EVERYTHING, especially the obvious stuff

“I know this is going to work!”

Of course every marketer thinks that, especially after reading some article about best practices.

Next thing you know, you’ve paused some of your old ads or implemented a big change on your website, without setting up any measurement test to confirm if it really works.

Not good.

Every time you change your email layout, add some new ad visuals or rewrite copy on your website, add to your to-do list a task to come back and check if it improved the results.

E.g. when I set up an A/B test on Facebook Ads Manager, I add to my calendar a note 2+ weeks ahead to return to the campaign and check the outcome.

Same goes about social media posts, emails, website changes, etc. – if you think a change will improve the results, make sure that it actually does.

Most of the time, you won’t probably see an improvement. What should you do in this case?

If the old and new version work on the same level (or there’s less than 2% difference) and you think the latest one better reflects your branding and messaging, keep it.
If there’s no big difference between the two options, keep the old version and start another A/B test to find another way to improve your results.

Another thing I’ve noticed: the bigger the difference between two tested A/B test variations, the more you will learn. Best to avoid A/B tests with very tiny variations in tested elements.

4. Key Learnings on 2018 about Facebook Ads

All the brands I worked with in 2018 run some Facebook Ad campaigns. It’s the easiest channel to get started with.

Also, having my Instagram and Facebook feeds flooded with ads by both local and global brands has been a perfect chance to learn what’s working.

Seriously – don’t underestimate scrolling your Instagram feed just to learn what other brands are doing.

Try this for 1 week, daily: Scroll through your Instagram feed. Take screenshots of all the ads you see. Put these to a G Drive folder or wherever you store your information. After 1 week, take a look at all the examples and write down the best practices you can use to improve your own ad visuals and copy.

You will end up with a personal gallery of Facebook ad ideas.

My Facebook ads folder has grown to 500+ snapshots

PS: If you’re interested, here are 142 Facebook ad examples I collected some time ago.

As this blog’s already full of Facebook ads best practices, I didn’t want to spend too much space listing new hacks & tips. Here’s a quick list of my new observations from 2018:

Facebook video ads don’t work – Yes, it’s an exaggeration. But with all the blogs and Facebook itself promoting the video ads to be the golden nugget of online advertising, I have yet to see video ads outperforming static ads. For sure, you should test if video ads work for you. Just make sure to measure their performance against single-image ads.

Prefer broad audiences – When setting up Facebook ad campaigns, leave your audience targeting as broad as possible. If you’re selling a consumer-focused product, you can even test targeting entire countries at a time – algorithms will deliver your ads to the most relevant audience.

3. Optimise your Facebook ad sets on right conversions – Whether you optimise your ad delivery on link clicks, signups, or purchases will have a huge impact on your results. If you have a high volume of down-the-funnel events, optimise your ad delivery on those instead of link clicks.

4. Use automatic placements – Yes, there’s an option to exclude some ad placement from your Facebook campaigns. However, it’s much more efficient to use auto-placements and, once again, let the algorithms decide what’s the best channel for showing your ads.

4. Use both illustrations and stock images – When it comes to ad visuals, it’s a good practice to use a mix of illustrations and stock images as different people relate to different images. Pro tip: using numbers in your ads will also help to improve results.

6. Create Lookalike Audiences – If you’re starting to advertise in a new country, it’s a very good idea to create a Custom Audience of your existing TOP customers all over the world, then create a local 5%-10% Lookalike Audience based on that. Simultaneously, set up a broad targeting ad set as it will at some point likely overperform the Lookalikes’ ad set (once you’ve reached all the people in the audience).

5. Value your time more and put 💸 into promoting each campaign

At Taxify, we do lots of fun marketing campaigns to engage with our riders and drivers.

Some of the most recent marketing campaigns include a Halloween giveaway, TOP drivers’ event in France, social media Advent Calendar, and more.

Some of my fave Taxify marketing campaigns from 2018

As you might imagine, a lot of time goes into planning and executing some of those campaigns.

But no matter how awesome those campaigns are, you’ll be missing out on a lot of potential if you just execute them without proper promotion.

What I mean by this is that when you’re doing a cool marketing campaign, e.g. a social media post in partnership with another brand, don’t stop by posting to your Facebook page. Make sure that you make enough noise for people to actually notice you campaign. You could do that by informing all your users via email, boosting your social media post for a higher budget, running online ads campaign to support the campaign, or use all of the aforementioned tactics combined. (And that’s so not the definitive list of things you can do to amplify your campaigns’ reach)

On a similar subject, don’t spend your time doing small marketing activities that have a very low reach and ROI.

Rather than execute 10 tiny social media campaigns or partnerships, focus on one bigger campaign that will truly engage your audience and will have a higher reach than all those 10 small marketing activities combined.

6. Create everything mobile first

How much of your work do you do on desktop? Around 100%, right?

This means that you view all the marketing campaigns, emails, websites, etc. that you’re working on from the desktop perspective.

Now, guess where your potential audience spends their time?

Mostly on mobile.

People read your emails on mobile, view your Facebook ads on mobile, browse your website on mobile…

Even this blog has a significant share of mobile traffic.

Desktop vs Mobile readers during a random time period

And you really need to remember the mobile-first approach every time you’re working on a marketing campaign. If you forget about it, here’s what happens:

Your email visuals that looked amazing in your desktop inbox are unreadable on mobile
Your website that was built looking at the desktop version is not fully mobile-responsive
Your online ads with copy in the image look overcrowded on mobile devices, resulting in low CTR and conversions

So really… 📍📍📍 Create for mobile first, and then make sure it also looks nice on desktop.

Cool articles, websites, blogs, books, podcasts and whatever

While it’s difficult to remember all the good stuff I’ve read in 2018, here are some of the nice books, articles, websites, blogs, books, podcasts I discovered.

Books:

How Brands Grow: What Marketers Don’t Know by Byron Sharp – this was recommended by a team member at Facebook’s Marketing Partners team as “the one book that all brand marketers have read.” I’m halfway through it, but it has been really informative this far
Sprint by Jake Knapp – I read it earlier, but the brainstorming sessions we did in 2018 based on the book’s suggestions all turned out to be amazingly productive
Ogilvy on Advertising by David Ogilvy – written in 1985, still highly relevant. Definitely worth your time more than any of the 2018 books on buyer personas and other random stuff

Most business books are similar in that they take 1-2 really good ideas that could be explained in a couple of pages… And turn those into a 200+ page book.

So what’s the point of reading these at all?

For me at least, the point of business books is not as much in learning a particular best practice. It’s more about the brainstorming that runs on the background while reading a well-written book. It kind of makes your mind wander and makes new ideas on how to improve your teamwork, marketing campaigns, etc. pop up – completely independently of what the book is teaching you.

Art Is Work by Milton Glaser – As explained by Amazon: “A lavishly illustrated, oversized retrospective of one of history’s most influential figures in international design evaluates the role of art in business today… documenting the creative processes of several of his works.” – note the part about “documenting the creative processes” – the book gives you a nice perspective on how to approach design work – any work for that matter
Joan Didion & Tom Wolfe – two writers who popularised the literary journalism movement throughout 60s and 80s. My favourite kind of fiction to read.

If you want to grow into a better marketer, reading just business books is not going to cut it. Differentiate your reading list with good fiction and art/design books to build your creative ideation and copywriting skills.

Newsletters & podcasts:

Benedict Evans, Andrew Chen, Next Draft (newsletters) – three main newsletters I actually open and read. I’m sure there are plenty of other good ones I’m not aware of tho. Please share!
Tim Ferriss (podcast) – not all of it is worth a listen, but check out the conversations with Seth Godin (another one), Nick Kokonas, and Naval Ravikant
The Design Better podcast by DesignBetter.co – I just got started with this one and while it’s about UX and UI design, a lot of the principles used in design are applicable to marketing and teamwork in general
Overtime Design podcast by Dribbble – If you feel like you’ve already heard everything discussed in marketing podcasts, switch to design podcasts

Online marketing resources and design galleries:

Facebook Active Ads (Facebook ads run by any brand) – In case you missed the train, Facebook added an option to see all the ads run by any advertiser in any country. Just go to your competitors’ Facebook pages and check out what ads they’re running. 👀
Really Good Emails (online email design gallery) – a super informative overview of the marketing emails sent by well-known brands
Land-book (online web design gallery) – I worked on two website design projects in 2018. Both times I returned to Land-Book for inspiration and latest UI hacks
Good Web Design (online web design gallery) – Another amazing gallery for anyone working on website design, This one’s more focused specific UI elements, e.g. CTA buttons, Pricing pages, and so on
Dribbble, Designspiration, Behance (online design galleries) – for fresh ideas whenever u run out of them

Blogs & magazines:

First Round Review (blog) – one of my favourite blogs featuring interview-style articles by managers and CEOs from top companies
The New Yorker, The Atlantic, (magazines) – if reading amazing reportages and essays from some of the best writers in the world is not a good enough reason for you, consider it an investment in perfecting your English copywriting skills. Best read on paper. 👌
The Purple Magazine, Frieze (magazines) – artsy magazines for great interviews (The Purple) and art inspiration (Frieze)
IdN Magazine – I spotted an issue of the magazine in a Paris museum bookstore, ordered a bunch of these when back at home, and found some nice design and campaign ideas. These magazines look sooo beautiful.
Trendhunter – I check this site way too rarely, but it’s a great go-to place for the latest cool stuff happening in branding

As you might have noticed, I didn’t read much marketing blogs. I’ve come to a conclusion it’s best to learn by doing.

Which is the perfect note to wrap up this year and this article.

But before that… the “I’d like to thank my mom” moment… To the marketing team at Taxify and guys at MeetFrank and MindTitan – I’ve had a truly great year together with you! 🙌