Spending £75,000 On Social Ads In A Month  — 5 Things I Learned

Armed with just one months practical experience managing big budgets on paid social, I found myself facing January in the travel industry. January is regarded as the most important month in travel. It’s the month with the highest consumer intent in the year so the industry makes a huge portion of its annual revenue in January. Ultimately, company performance in January dictates how difficult the rest of the year will be.

So, with a lot of pressure to figure things out quickly and simultaneously ensure the increase in ad-spend delivers a good ROI, I had to learn, fast!

Here’s 5 solid learnings I discovered after managing a £75,000 spend on Facebook & Instagram in a month:

1. Branded search plays a massive role in conversion

Personally, I think Facebook does an awesome job with attribution, it makes it really easy to see the lag between first interaction and purchase. You can even throw your offline CRM data into Facebook to be cross referenced with ad actions to show you how many offline transactions were influenced by your ad activity.

For example, if a user sees an ad but purchases over the phone because they needed support and guidance from a human, the email address of that customer (captured during the offline transaction) can be fed back into Facebook anonymously and then attributed to the ad that person saw. This was new to me so I thought this was mega!

Anyway, after playing around with attribution I realised that only 1% of the entire months purchases were being attributed to paid social on the last click. 1%! Maybe I shouldn’t be surprised by this, especially in the travel industry where it’s common for users to return 3–4 times during the consideration phase.

The majority of the last click conversions were attributed to organic and branded PPC. In fact, the more that was spent on paid social and the better ads performed, the more variance we saw in brand organic traffic such as an increase in queries like:

{BRAND NAME} + {PRODUCT NAME}

{PRODUCT NAME} + {BRAND NAME}

{BRAND NAME} + “REVIEWS”

Etc.

The most important take away from this was that brand SERPs (search engine results pages) should be as optimal as possible especially with a heavy paid media spend. So the following will be super influential:

• Setting up Google My Business

• Adding your best images and logo

• Illustrating legitimacy with plenty of detailed Google reviews

• Creating continuity with recognisable google post content

• Pushing down competitors found on your brand SERP with quality business listings, wikipedia page and independent review sites

• Control branded PPC messaging for continuity, trust and especially intent with call out extensions

2. You gotta go broad or go home

Early December we reworked pixel events into Google tag manager (rather than hard coded) and simultaneously worked in data layer arrays through our user journey along with an XML feed of inventory (if you run a legacy custom site you’ll know how satisfying that is to say). Spurred on by our Facebook account manager, we set up some catalogues and boldly experimented with dynamic travel ads and they rocked. But, they rocked even harder when I broadened the audiences.

This goes against everything Facebook historically encouraged which was to target your audience granularly. Now they encourage you to target your ads as broadly as possible to give the machine learning algorithms as much data as possible to work with.

It can feel really unsettling to effectively hand over your targeting to Facebook, especially with a large spend but… providing you’ve set the correct ad objective, campaign goals and bidding strategy, I would highly recommend you build the biggest but most relevant audience in your niche and set your prospecting loose, it works!

I have used the following audiences with strong results for a family travel company:

• Users aged between 23–54 interested in travel brands (pick 15 most relevant)

• Users aged between 23–54 interested in parenting brands (pick 15 most relevant)

• Users aged between 23–54 lookalike up to 10% on site visitors and purchasers (combined)

Focus your more granular targeting efforts on retargeting. Delivering more intricately to retain your prospects is arguably 6 times more important than getting another prospect. Why? Because on average it costs 6x more to convert a new visitor than a returning one.

3. Facebook data goes beyond your custom audiences

If you have a tendency to buy expensive things online or always browse at certain times or buy random stuff frequently, the pixel is watching you.

Even though Facebook has terminated it’s programme with third-party data providers that once fed its databases, it arguably has more data than ever before to aid targeting. Certainly more supportive data to target your ads than you think your custom audience might be capable of.

It’s rare to come across a popular website these days without a Facebook Pixel tracking your behaviour. The information Facebook captures in website data layer’s can build some immensely powerful profiles for targeting you as an eligible candidate for a specific ad when you jump on your social accounts.

The more relevant an ad is to you in your moment of discovery, whatever that may be, the higher the propensity you are likely to convert. So, Facebook’s pixel data feeds this aspiration to find users that are ready to convert ensure that advertisers keep spending.

Lets take travel to give you an example, here’s an ‘InitiateCheckout’ event that fires along side a data layer that sits in the source code of a big travel website’s ‘basket’ page (usually a booking summary page):

As well as the event generally providing a good indication that you are a user that is in-market for a holiday (because your on a holiday site and you’ve put something in a basket), it also collects the specifics on;

• A destination you like

• Your travel dates

• Your budget

• Even your party size

Facebook attaches this information to you as a user and uses it, not just for your benefit as an advertiser, and being able to target in-market users, but also for the benefit of the whole industry’s targeting ecosystem. Remember this is real-time data, Facebook is recording what you did today and combining it with what you did last week to serve you the most relevant ad.

This is why number 2 on this list is encouraged — the broader the audience the more chance you have as an advertiser to find the absolute sweet spot users who are in-market and ready to convert.

4. Direct/product focused ads were always favoured

“Ok, i’m prospecting so, people need to know about the brand, why it exist, the history, otherwise they won’t be interest…”

Not quite.

The brutal reality (which I should have anticipated because it’s totally how I behave with ads) is that people don’t give a flying plop about your brand until they’re invested in it, especially when you are interrupting their zombie scrolling on social.

Getting users invested requires their attention. Grabbing their attention requires ultra-specific content that is immediate and requires little to no cognition from the consumer other than ‘yes I like it and it’s relevant to me/someone close to me’.

I noticed this pretty quickly, only a handful of ads were hoovering up spend in-rotation. So, after reassessing, I tried to answer one question:

“What is the number one thing my market wants most and immediately”

I thought this question would give me the best outcome for my creative. Answering the question also helped create content that was specific enough to capture maximum attention. Attention magnet ads speak to a specific desire for the market, which in the travel industry’s case is usually awe-inspiring ‘holiday goals’.

In changing ad creative from being quite general with USPs and brand messaging to instead being ultra specific, price led, product featured and concise, I saw around a 30% drop in overall CPA, almost 50% drop in CPC and a whopping 60% increase in CTR which made spend go a lot further!

Bottom line: Serve ultra-specific creative if what you’re doing isn’t working and don’t forget number 3 on this list (your targeting is better than you think).

5. Best practice pixel setup can be flawed

Facebook will do it’s best to provide you with instructions to set up data layers and pixel placement however, my advice is to be scrutinous and challenge it. Generally speaking, the basic setup to track events that matter is as follows:

• PageView — All pages

• ViewContent — Product pages

• InitiateCheckout — Basket/Summary page

• AddPaymentInfo — Payment page

• Purchase — Thank you page

This setup isn’t going to fit with every business and every user journey. What if you don’t have a transactional website for example? Or what if you have multiple paths to conversion?

The main dilemma I experienced was one involving Facebook’s core machine learning algorithm. As i’m sure you are aware (if you’ve got down this far), your ads go through a ‘learning phase‘ which is an experimental phase Facebook uses to figure out which ad creatives, together with your audience, are most likely to deliver the best results. It then uses that learning to deliver the best results throughout the length of your campaign.

So, most of the time ‘conversions’ was the main ad objective I was using, and therefore the ‘purchase’ event should be the ideal campaign goal. But, getting out of the learning phase requires 50 conversions in a week. What if you can’t achieve that? Particularly for a small travel company who has online & offline sales to contend with, this isn’t possible even using ‘InitiateCheckout’ as a conversion goal.

So, the only option is to resort to ‘ViewContent’ as a conversion goal which jumps out of learning pretty quickly however, ‘ViewContent’ is only fired on product pages (if you’ve followed best practice).

This means, if you have any brand ads that send users to a homepage or a helpful category page, they will always be demoted because those URLs won’t fire the ‘ViewContent’ event immediately (triggering a conversion), whereas the product ads will fire the event straight away when users click through therefore demonstrating that the ad creative delivered a conversion.

Despite mentioning earlier in this post that specific and product-led ads are dominant, there’s still a place for brand awareness creative so, to make sure prospecting & retargeting has a mix of ad creative (i.e specific/direct and brand/general) you may need to add a ‘ViewContent’ event on your homepage and category pages.

Doing this for me at least, gave brand-led ads a fighting chance, diversified the content so it had more longevity and didn’t become stale and encouraged more users to explore what was on offer through more unique entrances to the website rather than being funnelled into one product page and bouncing.

Summary

The complexities of running successful paid social ads, at scale, are biblical. Personally, I love it because you can spot and tweak on the fly almost in realtime which can be really impactful. I also love it because there’s almost an element of inadvertent gamification with the way campaign setup & creative can be tested and tweaked to learn what works to squeeze the most out of your budget 🕹

And I think that’s the key to success with social ads, finding your own learnings. I actively avoided reading too much into best practice because I knew i’d learn more by ‘doing’ during my month of discovery. My advice is try as many things as possible; turn your creative upside down, target older people, build small audiences, build big ones, change your goals & objective, move your pixel events and don’t listen too much to the Guru’s.

Be your own guru.