Finlay Macgregor is a trainee paid search analyst, coming through the Katte & Co Fast-Track Academy. He gives us his take on his first six months getting to grips with paid search and explains the industry landscape and challenges he’s seen so far…

TLDR: “If you’re a ‘newbie’, you have an advantage. You don’t know the old ways. So get stuck in, you’ve got this, it’s complicated, and everyone is floundering, but keep your head up, and ears open, it’ll all be basic in a few months.”

I knew about advertising, Google Ads and paid search in a pretty vague sense six months ago when I started at Katté & Co. My closest direct experience was limited to Netflix’s “The Social Dilemma”, which I watched as ‘preparation’ for my interview day. Suffice to say it’s been a journey.

A huge shout-out goes to the team at Katté & Co, honestly, gems the lot of them. They’ve been super patient with me as I’ve learnt the ropes. Now though, I’m finding my feet a little, this won’t last long I’m sure, but while it does, I thought I’d share some thoughts on the landscape, from a ‘newbie’s perspective. Fresh, naive, eyes I admit, but hopefully something that will be useful to others.

I started with the Google certifications, which were great, but there is some updating needed, I mean broad match modifiers don’t exist anymore, that’s a pretty big one.

What I left those thinking was, but what do I do, and where am I? The actual day-to-day stuff wasn’t clear, nor was the landscape and current situation as a whole. That day-to-day stuff comes from getting stuck in, but the landscape – that takes a bit longer.

Now, I’d say there are two main parts to the landscape: 1. cookies are dying; 2. automation and machine learning. After that, there are some other, less in your face, bits: GDPR’s effect on first-party data, and the value of YouTube & social media.

Cookies are dying:

Due to changes in iOS and GDPR among other things, people don’t like, and don’t have, cookies to the same extent they used to, and this is on a downward trend. As a result, tracking the conversions we so love becomes pretty tricky. We can track when they happen, we can tell when files are downloaded, or items are bought, but how do you attribute that to an ad click if tracking isn’t a thing?

Well, either make tracking a thing or be clever about it. Google being Google, they’re going for both. Consent Mode and Enhanced Conversions (working together!) let this happen. I could try and explain these in a blog post, but as much as I know now, Google and others have a much better set of explanations. Suffice to say:

Consent Mode: Makes Google’s tracking adapt to how the user chooses cookies on the website banner – if they say yes, nothing changes. If they say no, then the URL tag is used and we rely on Google’s machine learning conversion modelling. Basically though, cookies are not used. See the Google Tag Manager god Simo Ahava’s explanation for more technical detail here.

Enhanced Conversions: Google’s machine learning uses first-party data to match against converter’s Google Accounts info (all anonymously). This helps as it tracks people across sessions to get a more ‘enhanced’ sense of what your ads are actually leading to. DataFeed Watch have a nice explanation of this here.

So Google has solutions (yay!), but as you can imagine, this is still a paradigm shift in how advertising and online tracking works (boo!). To combat this, plus the age of the existing platform, Google have rapidly updated their analytics platform to Google Analytics 4. From July next year, the current platform GA3 will die, and if you want year-on-year comparisons, you need to get GA4 at least collecting data by July this year. This is just a fact – and a scary, looming one at that.

So it’s all change, change, change! So far: Cookies are dying, so Google is leaning into clever workarounds, first-party data and machine learning to sort this all out, and their analytics platform is changing too to keep up, and we need to match pace with that.

Machine learning – Because humans are wayyy too complicated for humans to understand:

The paid search landscape doesn’t stop there. Humans are complicated. In writing this I used at least three different computers/phones, two major operating systems and three word processing packages. And that’s just a blog post – imagine what buying something is like! Meagre humans can’t keep track of and advertise across all of this. As a result, Google pushes for machine learning solutions – enter Performance Max (& Co).

Much like GA3-4, Google are forcing people to Performance Max (an even more widespread and machine learning-driven version of smart shopping campaigns) this year. With Broad Match and responsive ad formats becoming the standard, if not a little old hat by now.

So machine learning is a huge deal, and we need to learn to manage it, work with it, and know where to stop it. As much as we can be impressed by machine learning and be amazed when it drives increased performance, if you want neat traffic with broad match, you need a lot of negative keywords. If you want nice placements and ads from your Performance Max or Responsive Display Ad campaigns, you’ll need to make nice and clever creatives. Overall machine learning is great, but it isn’t as neat as people need or want yet, and potentially never will be. So we need to keep an eye on it, but also acknowledge (with a healthy dose of humility) where it can do FAR better than we ever could.

First-Party Data – sadly not really a party, more a compulsory lecture:

Mailing lists, client lists, customer databases – all these things let you have first-party data. At a basic level owning data is now a big deal. Data was always valuable, but GDPR makes owning it (the right way!) essential.

If you own first-party data, you know that it is real, and you can do things with it. So start accepting you’ll have to make accounts for everything you do online, because it is how people keep track of you. From purchase history and age, to location and everything else in between. For us, understanding how to use first-party data in advertising, from customer lists to remarketing, is now more important than ever.

Machine learning can work with the zillions of tiny signals we can still track to work stuff out, but to personalise ads to the extent we are used to, in the creative, human way we like – that needs very high quality, and information-rich, first-party data sets for us and machine learning to build upon.

YouTube & social – come on, you knew it was coming:

With all this talk of how complicated and important humans are, it’s no wonder the most direct and potentially provocative platforms on the internet are also pretty important. I’m kind of lumping this all together because there are a lot of similarities. These platforms and placements provide a more direct, or arguably meaningful way to communicate with people at a higher level, with more video, audio and images.

Given this, we need to target them precisely, and in the best possible way. Requiring, unsurprisingly, first-party data to work out who to target, and machine learning to work out what works best across all the signals the platforms can collect.

Thankfully, there are tools out there to make this easier. So, follow the best practices, work out how these platforms can work for you. And even if it doesn’t seem like they will work, try them. These are new avenues, at least in their current forms; newbies learn fast, so try and learn the newest stuff too, it’ll put you in good stead.

In summary:

  • Cookies are dying.
  • Google are using workarounds, first-party data and machine learning to get around this.
  • They are also transitioning at insane speed to a new analytics platform, and woe betide anyone who falls behind.
  • We need to get used to machine learning because it is here to stay, and that means knowing how to limit it.
  • First-party data is becoming the only source of definitive personalised information we have to base ads around. So we need to know what it can do.
  • YouTube and social media aren’t just to be mentioned in passing – they’re invaluable, so get stuck in.

Conclusion – FINALLY:

This seems like a lot, but if you’re a newbie, you have an advantage. You don’t know the old ways, GA4 will be your bread and Performance Max your butter. Machine learning will be the old mate you’ve known since the start, this will all be natural to you. So get stuck in, you’ve got this, it’s complicated, and everyone is floundering, but keep your head up, and ears open, it’ll all be basic in a few months.

OK that’s it now, good luck, and don’t let machine learning run all your advertising – yet!