Optimise Ads For Big Returns
Introduction
So far in our series on Creating Value From Data, we have focused on the strategy and the tools that you will need in place. Once you have the systems up and running , the next biggest question is of course, what to optimise? Amidst the myriad of demands and priorities of different departments, pinpointing where to allocate resources can be challenging but one area which has consistently stood out in our client work has been optimising digital advertising.
A Retrospective On Digital Competitiveness
In order to frame our problem, we must first outline the history of the digital landscape to highlight just how competitive it has become. As you may no doubt be aware, in 1991 there was only one website in the world. In 2001 there were 29,254,37 websites. In 2011 there were 346,004,403 websites, and in 2021 there were 1,930,614,923 websites. That means that in the last ten years alone there has been a 487% rise in competition.
As more websites have been created, so too has the demand for advertising grown. In 2009 there was major news that the keyword ‘mesophilioma’ sold on Google for an incredible $99.44, the most expensive paid click sold ever at the time. In 2023 the most expensive keywords on average was ‘offshore accident lawyer’. And the cost on average (meaning not the highest) was $815. So that is more than 8 times more expensive in a 15 year time period. These broad stats give us a nice outline of just how much the cost of online advertising is going up, fuelled by increased competition and the likes of Google and Meta having a stranglehold on the industry.
Should You Trust The Algorithm?
A common dilemma for marketers today is that, due to consumer privacy concerns, they are becoming more restricted in who and how they can target people. They are seeing their costs soar, and being told that they should go broad and trust the 100% automated advertising campaigns such as Performance Max to find their audience.
The problems with this approach are:
No Competitive Differentiator: If every business in a market relies on the same algorithm for their ads, they will all be optimised to the same customer profiles and so instantly lose competitive advantage outside of increasing their budget and the quality of their landing page. This emphasis on cost means that spending power becomes the most important differentiator in winning customers, which is exactly what the advertisers want!
It still takes time: Despite the promises of rapid advancements in machine learning and artificial intelligence, training a sophisticated algorithm remains a time-intensive process. It can certainly take a few weeks in order for tools like PMax to optimise towards your audience. This learning period often frustrates marketers and leads to a level of mis-trust in wondering just how much budget should be spent before the algorithm will deliver the results.
Greedy Advertisers: Conversion often isn’t linear and someone might click on multiple ads served by multiple providers before converting. This is why you can have a scenario in which you have 2 conversions but only one paying customer, as both platforms will claim they were responsible. Essentially meaning that without data to understand which action had the biggest sway, you end up paying twice.
Optimised towards conversion: Quite often, the primary goal of and ad channels’ algorithm is to maximise conversions. Conversions are often set up by the user themselves but tend to include actions such as forms being filled, items being checked out or sales.This focus on conversion rates can drive short-term revenue and demonstrate immediate ROI, but it often fails to capture the broader picture of customer relationships over their entire lifecycle.
Conversions vs Lifetime Value
Conversions represent a transactional view of customers. Conversions from new customers will typically represent a smaller $ value when compared to their purchases over their lifetime – as trust will need to be built with you. If you are offering some sort of initial deal through advertising, customers may take this up and then never shop with you again! All of this to say that a first interaction with your brand is not reflective of the longer relationship, and so this is why understanding Lifetime Value is so important.
So to give you an example; Emily clicks on your website today. She will probably buy something quite small for £20. But over time if she trusts your brand, her lifetime value could be a lot higher, say £2000. And so rather than optimising towards the first click or X amounts of £20, you should optimise towards how many £2,000 you can get.
Even at a very basic level if you figure out your Average Purchase Value (total revenue divided by total purchases) and then Average Purchase Frequency (number of purchases divided by number of customers) your Customer Value will be your Average Purchase Value times by your Average Purchase Frequency.
You would then figure out your Average Customer Lifespan by dividing your total Customer Lifespans by your Number of Customers.
And so Lifetime Value = Customer Value x Average Customer Lifespan
Even when you are a new business who might not fully understand the lifespan of your customer, average order value can help to change your thinking on marketing spend, and of course, where you have the capability you can make this model much more sophisticated.
Using Your LTV Model
Once you have your LTV modelled, it will normally live inside of your data warehouse – so how do then use it to optimise ads? You need to send this information back to google/meta etc using Reverse ETL.
Reverse ETL takes modelled data and pipes it back inside of applications that your teams use. In this case it would send the expected LTV values to Google and Facebook Ads as conversion values, thereby optimising the campaign that generated the lead. This approach does not rely on user browsers or cookies, ensuring a higher quality match over time. We will talk in greater detail about activating data in a later blog.
Real-World Comparisons
173tech have used LTV to optimise ads for a few different clients, we wanted to include them here just to demonstrate the impact, no matter your type of business or advertising budget. In all of these cases the cost of acquisition was significantly lowered, giving us our savings number.
Washclub ( A pickup and delivery laundry service ) Initial marketing spend $86k – Saving 72% – Monetary Value $62k
Elopage ( A SaaS-Tool made for entrepreneurs to manage their digital businesses ) Initial marketing spend $400k – Saving 36% – Monetary Value $144k
Momentary Ink ( A D2C temporary tattoo company ) Initial marketing spend $2m – Saving 12.5% – Monetary Value $250k
Petlab ( A D2C pet food company ) Initial marketing spend $8.75m – Saving 40% – Monetary Value $3.5m
What these numbers mean in real life is that for the same budget, these brands can attract more customers. In Washclub’s case literally three times as many customers for the same cost. It is also important to note that the savings outlined here only compare to the previous year, while ever the model is running, these companies are not only saving money in lowering their CAC but also attracting customers who are more likely to spend higher amounts with them.
Conclusion
Whilst companies will always have conflicting priorities, optimising your advertising is a great place to demonstrate the value of data to your company. It has a clear impact on the bottom line and acquisition is an area that every business is looking to improve upon. Clients typically save 20% and up by doing this and so if your advertising budget is above $75k then this work will quickly pay for itself!
Get in touch with 173tech to see how we can help you in this area.
More Articles
Our eBooks
Our Services
Contact Us