Performance marketing is broken

Watch the webinar and read the highlights below



There is always something changing in the Digital Marketing space. Whether it’s new platforms, how to best track traffic and conversions, or how individual platforms’ algorithms work; and now with changes such as third party cookies becoming a thing of the past, and more and more machine learning being incorporated there is even more to stay on top of. The current ways of doing performance marketing need to adjust again but don’t worry, Booster Box is here to help.

About Booster Box
Booster Box is the best performance agency on the planet*. Based in sunny Tuscany, Italy, we are a team of Data Scientists, Mathematicians, Developers and PPC Specialists. (*according to our Moms).

They’re also an awesome customer of Funnel’s!



Check out some of the highlights from our webinar with Booster Box’s Gianluca! And if you’d like to download the presentation, feel free to do so here.

Why (if you’re an eCommerce) you should move from ROAS to Future Margin on Ad Spend

ROAS [Return On Ad Spend] is sort of like the evil twin of ROI (Return on Investment) is a broken metric. And it doesn’t make any sense that we’re optimizing our Paid Performance campaigns with ROAS as a KPI, as a north star. Why is ROAS a broken metric, and what can we do about that? There are really two reasons why ROAS is broken:

  • In ROAS we’re looking at revenue, and not margin. If we’re drawing a comparison between the financial world, it’s just like placing our bet on a stock, just looking at the revenue and not profit...clearly, when it comes to eCommerce, margins are telling a much more complete story.
  • ROAS doesn’t include future transactions. How often do we as Performance Marketers split our brain in 2? On one side we’re all focused on getting people to buy for the first time, on our eCommerce or our lead gen, and the other half of our activity is to make sure that people are coming back. We’re doing that with remarketing, email marketing, marketing automation, with all sorts of ways to keep getting in touch with customers and having them come back.
    But often, when we go and optimize for ROAS, we are excluding the future transaction and the future LTV [Lifetime Value] that we are getting from a future customer. So that is obviously giving us only a partial view of what we could do.

[Instead, you can move to measuring] Future Margin on Ad Spend. Basically, we’re taking into consideration what are the future transactions that we’re getting from a customer that we’re willing to acquire, and then what is the expected margin that we are expecting to generate from this customer in a given span of time.

…[And] I want to take a minute to underline how much this change will give you a much, much clearer view. It’s like us and our competitors, we’re looking at a bunch of stairs and there’s fog, and everybody is looking and can only see the next 5 steps. [But] if we’re moving to this metric [Future Margin on Ad Spend] suddenly we can see for the next 50 steps. So it is not perfect because there is a lot of assumption on how many future transactions we will get, and what exactly the margin will be for future transactions, but it will definitely be a much clearer view than the competition.

So it means that we can evaluate every single bid, every single placement, every single creative in a much more accurate way.

Step to take: have the conversation about redefining success (especially for eCommerce) about how to define success and start moving to a version of using Future Margin on Ad Spend instead of ROAS.

Algorithms will be a commodity, so feed them with the right kind of data

...we’re all using the same algorithms. In other words, the algorithms are leveling down the competition. Advertiser A is using algorithm A, Advertiser B is using algorithm B, but they can both switch to the other guy’s algorithm in a click. So, what is really our competitive advantage? It’s not exactly which bidding model we’re using or what ad testing engine we’re using, it’s not really the algorithm itself - the algorithms are becoming a commodity - what is really a competition factor, what is really the secret sauce, is the data that we’re using to feed our algorithms.

So, I think, as well, that this is a fantastic opportunity because most of the world is just blindly in love with the idea of using a smarter algorithm - ‘let me test something that is bringing more performance’ - but that’s not really the point. In 2 years' time, algorithms will be a commodity and we all recognize that it’s really just a formula. It’s a fancy way to say formula, but we all recognize that the math behind that is empowered, actually, by the data that we’re feeding it with.

Step to take: Connect your audience data to the platforms so that you can start feeding your machine-learning algorithms with some good first-party data. Gianluca provides a checklist to get started doing this.

A 7 step checklist

  1. 1. Get the basics right: UTMs, GTM
  2. 2. Pick a database: BigQuery would do
  3. 3. Connect CRM data into your database (Funnel can help)
  4. 4. Connect Ad Platforms data into your database (Funnel can help)
  5. 5. Define your attribution model
  6. 6. Define your clustering
  7. 7. Push back your audience and LTV data into the Advertising Platforms (eg API Connectors)

Why you need to think in terms of people, instead of keywords

We need to think of people, not keywords. And the platforms, particularly Google, are really catching up in this sense - meaning that platforms are moving more and more towards audiences. So to push us to think more in terms of audiences, rather in terms of keywords.

...I’m not sure if audiences are the new keywords or not, that probably sounds like a big click-bait type of article, but definitely, audiences are becoming more important. So that is something that is [also presenting a lot of opportunities] for us because now we can now...empower our campaigns even more when it comes to Google and Facebook.

Why you need to have an agnostic, aggregated view of your advertising data

The reality is that platforms are operating in silos. So they don’t talk to each other and it’s very very difficult for us as an advertiser to have one unified view that doesn’t have [a platform] that has some sort of skin in the game - that is agnostic enough and will allow us to make the best decision for allocating our investments.

...Right now, it’s not only about if we can trust, 100%, [the ad platforms], the question is more complex. Meaning, we don’t have, right now, a full way of monitoring the user journey across different platforms and across different devices and we don’t have one single source of truth because we’re relying [mostly] on the data we’re getting [directly and only from the ad platforms].

So in all these discrepancies, in all these inconsistencies, lies this massive opportunity for optimization. So bridging the gap between different platforms in one aggregated, agnostic view is a fantastic driver for optimization.

This is where Funnel can help ;)

Cookie-Restrictions will require marketers to adjust their way of working

Cookie restrictions will change the game for us, and that’s something that we need to take into consideration because we need to move more to first-party data - we cannot rely on cookies and tracking systems as we know them today.

[40:48] Will cookies be restricted on ALL platforms?
The restriction that will happen at a browser level, will probably be the main challenge here. So aside from - I don’t know the specific policies or the product roadmap [or plans the other platforms] have, but the data Google Chrome decides to switch off that is a game-changer for all the other platforms out there.

To go back to the Big Short analogy and back to the financial market [analogy], if you go back to the day when Google announced they were going to [restrict cookies] and you look at Criteo’s stock, you will see a decline, right? So the financial markets already decided what is the outcome of the action that Google is enforcing.

The good news is that it looks like we have a couple of years to get ready for that. And the even better news - companies, we know our clients, we know what they want, we know their taste, we know our products, right? And we know a lot of the information that’s extremely valuable around and inside our company or our client’s companies. This is information that is sitting somewhere in an excel file, in a CSV file on someone’s computer that is unused...All of the emails that were collected in that shop in London once we ran that promotion...all the dusty and old CRM data that one day we’re going to use to do something more than email, but maybe one day - that day is now. Now is the time where we can take all the knowledge, everything we know about our customers, and try to use that information to feed Google, Facebook, etc. with first-party data. And that will allow us to overcome the cookie restrictions - partially, to overcome some of the cookie restrictions coming our way.

Step to take: Start gathering the first-party data you have and find a way to feed it into your advertising platforms to help partially overcome the cookie restrictions coming our way.



Gianluca Binelli, Founder & Managing Director, Booster Box

Gianluca is the founder of Booster Box, a performance agency specialized in scientific marketing. Prior to Booster Box Gianluca was at Google for 6 years working on different teams within AdWords. First as a Product Specialist, he then went on to manage online marketing for Google’s products in EMEA. Gianluca also served as Capital G's Advisor, helping start-ups with their online marketing in Google’s own late stage investment fund. He is a regular speaker at international PPC conferences (eg AdWorld Experience, HeroConf, Web Marketing Festival). Gianluca teaches digital marketing at the University of Pisa. In 2019 Gianluca was voted 2nd Most Influential PPC Expert by PPC Hero.


Megan Lozicki, Content Marketing Manager, Funnel

Megan has been at Funnel for 3 years, starting on the Customer Success team and now focusing on the marketing side of things.