Searching for answers. Navigating the new digital maze to deliver results.

Working in digital performance and specialising in Search optimisation for over a decade, I thought I’d seen it all. But when we began working with the University of Melbourne on creating an optimised search platform for their entire student recruitment program, I realised our understanding of what makes Search work – and why it has become so critical to prioritise it in our new business environment – was evolving right in front of me.

Somewhat surprisingly, the answer lay in an economic principle first put forward in 1975, well before Google, Search Engines, Social Media or, quite frankly, the internet.

Working out Search Advertising is essentially a maze and the quickest way to solve one is to invert it. “Instead of looking at exactly what we want we need to just look at exactly what we don’t want.” – Brendan Keevers


“When a measure becomes a target, it ceases to be a good measure.”

In other words, when we set one specific goal, people will tend to optimize for that objective regardless of the consequences. In the case of search optimisation, if you declare search itself as a primary metric, you’ll definitely be at risk of Goodhart’s Law, right? Well, not necessarily – it all depends on how you define success.


In the case of University of Melbourne, it seemed on the surface pretty simple: How many new students can we recruit? However, to have a successful search driven platform across the multiple types of courses and program offered at a University, we need to go much deeper.

By declaring objectives within search, particularly using Google Ads for SEM, you need to work out how your full conversion path works both prior to search – and also once traffic reaches your website. A great way to review this is understanding how many clicks or interactions the user has to make either side of your search medium in order for “the sale” (in this case a live recruit) to happen. These we came to call “Soft Conversions” (prior to the Ad Click) and “Hard Conversions” (after the ad click but before the sign up).

Then we start to look what is happening before prospects reach search. Because Search Ads has a native ability to learn through AI and Machine learning, if we can find out what drives Hard Conversion, we can trace their behaviour all the way back to where they began the process. We add to this around a hexabyte of statistical information, letting us match the user behaviour and power the bidding process. If the user looks to be following a similar path to what converts, we actually don’t need to be that specific on targeting. All of which means we can reach far more people more often.


Finally, we look at the actual nature of search and how we can optimise it. But to do this we need to understand the Three States of Search:

  1. Brand: People are actually looking for you.
  2. Product: People are searching for your specific products or services.
  3. Generic: People are searching with your category. These are brand agnostic and general solutions to your customer’s needs.

Once you can split your Search ads approach into these 3 main categories (with some fringe exceptions) you can start to understand the volumes and demands for each. By mapping this and then tracking the user through these spaces we can optimise the best pathways along each. We then make sure our ads have the best ranking possible as SEO and User Experience (UX) will drive the best ad performance and lower your CPC (Cost per Click).


Now, how these stack together gives you your best average performance across your offerings available on Search. This still links towards your main objective of the hard conversion, but by splitting it into 3 different categories, you then reach users in three different ways based on how they’re searching.

Once we can understand this as a common denominator (considering seasonality and other forces) we can start to look at the statistics that Search can provide us that no other paid media can, the action results.

But after all that – does this help us solve Goodhart’s Law? It’s quite simple really: We must look at it like a reverse maze, as the behavioural data and search intent does. Instead of looking at exactly what we want we need to just look at exactly what we don’t want.

Once you can begin to define your key exclusions you can use search, your behavioural intent and also how you shape your conversions into a Guess Who board of Search Engine Marketing. Not only does this work on the notion that saving money is just as good as making money (as you’re looking at it from a standpoint of not spending on non-converters), you’re eliminating the specifics of what you want and remove yourself from the confines of Goodhart’s Law and the issues that stem from it.

So, does it work? Well, over the 2 years that we worked with the University of Melbourne we had an overall growth of completed applications through SEM of +1,493%. To achieve this we ran almost 400 different campaigns with around 2000 different ad groups covering 10 different faculties.

As you can see, it’s not easy – but it is worth it. And in the ever-changing digital landscape – it’s never been more important to get search right.

PS… If you want to know more about Goodhart’s Law, check out this excellent article by Will Koehrsen:

If you would like more information about this, please reach out and we can provide a detailed discussion on how we can help you with your Digital Marketing approaches.

SMART is ready.

We love a good business problem. If you'd like us to take a look at yours, 
you can ping us anytime at