“Half the money I spend on advertising is wasted; the trouble is I don’t know which half.” John Wanamaker (1838-1922). Are we any closer to answering that fully in 2017?
Calculating CPA (cost per acquisition, also referred to as CPL – cost per lead) is a challenge. But tracking the cost of your first acquisition point (which may be a lead or a customer) is essential in understanding marketing channel efficiency and equally marketing-mix efficiency.
Every business for decades has struggled with two common things: which channels to focus marketing spend on and how much return every dollar spent returns. This isn’t as simple a question as it may initially look.
Marketing channels have evolved hugely since television, radio and print ads were key ways to reach a mainstream audience. Despite the advancement in tracking technology and ad platforms, the question of how best to utilise each marketing dollar remains a challenge, even for the most advanced marketing operations and platforms (despite what they tell you) and here’s why.
Scenario 1 – a customer sees an ad in a single channel and purchases there and then
CPA challenge: ensuring there is tracking in place, which for digital is usually in the form of digital analytics and for offline is primarily in the form of promotion code.
How often does this scenario occur? There will be a good percentage of your acquisitions that only ever see your one ad on one platform or through a single medium and decide to buy all within the same ‘session’. For clarity we define a session here as a stream of events all happening one after the other.
For simplicity I’m focussing on a visit to a website as the journey to complete a transaction (as opposed to a shop visit or app download via an app store). So:
- Customer sees ad
- Customer clicks/responds to ad
- Customer visits website and adds item to cart
- Customer buys item
In this scenario, the customer does this without a break to browse other sites or coming back later. For a business with limited brand and product awareness, it is less likely that someone will see an ad for the first time, click and buy in one go.
Of course, we marketers would love to think our ads are spot on and the target audience have been reached in one hit and that audience is in the mindset to buy and they visit your website and buy without even thinking about other important things in their day like what’s for dinner or the latest cat gif on Facebook.
There are also factors around trust which affect consumer behaviour – trusting how legit the company and website is, whether the product quality can be trusted, whether the product price is going to be the best you can get etc.
What kind of company has enough brand awareness and product range association that seeing a single ad will prompt you to both trust the business and purchase in one go? Amazon comes to mind as a company that can do this, as may other well established businesses, where the consumer is fully trusting of the brand and broadly aware of the product range. But how quickly do you need to increase sales? And how long have Amazon taken to establish themselves?
Trusting that a company like Amazon aren’t going to defraud you, are going to deliver your product, and are going to deliver a known quality – these are questions which for most people are answered on the Amazon site or through strong brand association. You, the smaller brand, simply can’t match that.
Chances are that if your data suggests that a given transaction occurred from the first ad the customer saw in the same session, there’s something else at play. Let’s look at scenario 2.
Scenario 2 – a customer sees an ad, doesn’t interact, but comes back later to buy from you via the website directly or via another marketing channel.
CPA challenge: tracking the attribution of the source channel due to the lack of connection to the final conversion channel.
This is a very common (if not the most common) scenario and essentially how brand-led campaigns make their mark by creating an uplift on trackable channels. In this scenario a customer will see an ad, for example on TV, outdoor display, a leaflet, print ad or even digitally such as Facebook or Youtube as follows:
- Customer sees ad
- Customer doesn’t interact with ad
- Customer visits website directly at a later time/date and adds item to cart
- Customer buys item. Ka-ching!
If you’ve ever run or been involved in a brand campaign you’ll know what a struggle it can be to truly measure the impact of the campaign on bottom line numbers. As a performance marketer you’ll know how fluffy ‘brand awareness’ can seem as a metric.
To fully measure brand impact requires many more tools such as a brand impact test which measures trackable channel performance with and without the brand campaign, baselining your organic/non-paid channels to measure uplift and things like paid search brand keyword uplift in particular.
That being said, not every ‘brand campaign’ is big budget, massive exposure stuff. Every supermarket carrier bag shows branding at play. Every time you read a newspaper with the logo facing outward branding is at play. Branding plays a big part in our everyday interactions with companies, whether we notice it or not. And as a marketer, the impact of brand awareness is often neglected when calculating the ROI of your dollar spend.
If you’ve scaled a business, you’ll know the importance brand plays in creating the foundation for sustainable growth. Sustainable growth is that which doesn’t rely on any single marketing channel (eg Facebook ads) and has a good blend of non-paid acquisition acquisition channels.
Scaling a business solely focussed on the hard metrics of spend, sales and CPA will hide the subliminal impact your marketing is having, like how Pavlov had his dogs salivating at the sound of a bell, or how those that have dined there, have an instinctive, uncontrollable reaction to seeing the Golden Arches through the release of dopamine.
Find me a direct response ad that can release dopamine better than a well placed brand campaign. Are you tracking the DU (Dopamine Uplift) of your marketing campaigns? Surely this will trim your CPA by a percentage or two, yet we’re often having to make budgeting decisions based on the reaction of the rational brain when neuroscience studies confirm the unconscious, deep thinking part of your brain as being a key driver to decision making, including what and where we buy.
The brand funnel is about creating awareness in defining your ‘who’ and ‘why’ – who you are and who you are relevant is what creates meaning and connection to your target audience. Too often, businesses focus on what they do, completely bypassing the customer need. This is a typical sales-led approach to growth and not a customer-centric view.
It is the channel CPA-centric view that drives marketers and businesses to focus heavily on what they do and this isn’t what your target audience will primarily care about. Give them a reason to not just be interested, or to like you, but to love you as a brand and watch those CPAs fall.
Below is a good illustration of the brand journey and impact on your acquisition funnel – moving from a wide audience, some of which are not your target consumer, and funneling those that are toward your business.
As a business, if your audience is everyone, then whilst your product may well cater for every single person you can market to, until you create relevance for that person you’ll struggle to help them determine how your product improves or complements their life. This is the role of brand marketing, to sift out non-consumers and attract potential consumers from a wide funnel of prospects.
The ‘indifference’ stage noted in the diagram refers to the fact these people have no positive or negative opinion on the business.
The relevance of this user journey scenario is this: whilst you may measure and try to improve acquisitions that come through to the website or through brand searches, do you really know what the first interaction for that customer was? If not, then deciding where to place that next dollar is not going to be entirely fact or numeric based.
As a Chief Marketing Officer your growth decisions need to rely on a variety of datasets to analyse impact, performance and future actions.
Using multi-data points for indirect attribution analysis
An important and often used piece of attribution analysis is a question asked to customers after the point of purchase: “how did you hear about your us?”.
Don’t underestimate the importance of this simple question. Although we have the ability online to track the channel from which a visitor landed on a website (eg search engine, facebook, youtube, affiliate etc through web analytics), this question gives us a valuable first-touch data-source for channels which we can’t track so well such as TV and print media.
In the case of ecommerce companies, many rely on promotion codes in order to track the marketing source but in this scenario, where a customer has seen an ad in one source and purchased via another source, you’ve lost the originator marketing channel.
For example if you distribute leaflets and a customer sees one, then goes online and purchases through an affiliate code, you have no concept of the performance of the leaflet and may make the incorrect decision to stop distribution of inserts due to low uptake.
At toucanBox we use a post-transaction survey to ask customers how they heard about us at the point at which we are front of mind. The reason to have this in the checkout confirmation page and not in the confirmation email or later is because it would otherwise only serve to dilute the quality of response.
The positive impact brand awareness has is best demonstrated when we look at TV advertising performance. We have a certain CPA target we are aiming for and use a promotion code on the TV ad to help us track customers coming directly via the TV ad. However, if we follow scenario 1 above and assume all customers that see the TV ad, will come directly to the website, our CPA looks quite astronomical and unprofitable. Decision? Let’s trash TV it clearly doesn’t work! Ah but if you add additional data points, it tells a very different story.
If we look at just one of our many paid search keywords related to brand searches (see below), we see that there was a rise and fall of searches on that particular keyword when we ran TV advertising in February, denoted by the rise, peak and fall which correlates with TV spot timings. This gives as an additional attribution back to TV because the increase in paid search hits for this one keyword can be directly matched back.
We analysed all brand keywords that we could track, alongside those people directly visiting the website and those coming through non-paid search and found similar patterns. The uplift in visits also correlates with an uplift in transactions confirming both volume and quality of TV impact for this keyword.
Now if you tally up the additional sales TV brought in as the originator channel into the trackable channels (brand paid search, organic search and direct visits) the combined CPA comes in within target. The correct decision therefore based on this data is to continue with TV advertising. I may not know the complete, numerical positive impact TV has across all channels but I can be convinced with this data that it is worthwhile.
Another data point we used was the post-transaction survey mentioned earlier. In response to ‘how did you hear about us’ we gave some relevant options such as TV, print, search etc. When analysing the data for the same period, the majority selected ‘TV’, which was of course great news but unsurprising.
However where this was even more interesting was matching this back into Google Analytics as an event and analysing those that said they heard of toucanBox via TV and the final channel they converted on. Through this we also found TV impacted sign ups that came via Facebook ads, online partnerships and affiliates.
Essentially we added a layer that analytics tools cannot yet add – where did they remember seeing or hearing about the brand first? This is important because we’re looking for the trigger ad that prompted them to move from the indifferent stage to the like stage for the brand.
TV for toucanBox had played an even bigger part in uplifting the entire marketing mix. Without the additional data points, a singular view of CPA and attribution would have led us to making the wrong decision and hindered growth.
No surprise, CPA tracking is still an issue
Despite this level of insight and understanding, can we answer how best to place every single marketing dollar?
No we can’t, because the true CPA of every channel, considering what percentage of the customer journey each channel contributed, is a thing of data science beyond the scope of this article. You can search for and sift through tens of thousands of articles written on this subject in the last 5-10 years yet we’re still not able to 100% define a true CPA.
The challenge is that typical tracking of budget and transactions, from channel to channel, device to device, across web analytics and crucially your customer database rarely match up. This occurs because of how each tracking system sees the view of a single transaction.
If a customer visits the site from a Facebook ad, based on the attribution window you are looking at Facebook will record it as a transaction via Facebook, regardless of whether that person converts subsequently via another channel or not. Analytics tools such as Google Analytics register the last channel the customer came through and although GA offers a way of viewing multi-channel attribution, at best some marketers and analysts are still inexperienced with how to use it for budget decision making and at worst are not even aware it exists.
GA will also often under report against what your customer database shows due to cookie/tracking issues and so as a marketer it is a constant struggle to optimise channels, marketing-mix and internal KPI reporting. The screengrab below is a typical view of what our customer journey might look like from first click to last. Quite the journey. And if last click was all we went on, we might ditch email or organic search from our marketing mix.
Whilst at toucanBox we are highly data-driven in our marketing analysis and decision making, we’re also in fast-growth stage and so require a balance between data accuracy and growth agility which often increase inversely. Making decisions on hard, measurable data points isn’t enough to make a fully qualified decision to assigning the true CPA for a given campaign or a channel.
So after all of that, how exactly do you evaluate the return on your dollar spend?
The simple answer is through complex data modelling, tracking and attribution. Well no, actually the simpler answer is you don’t.
Instead marketers will often take stock of a blended view of marketing in order to determine the success of a campaign. If tracking and accounting the return from every dollar is your priority (and there’s no reason it shouldn’t be) then you will need to have the time and resource available to create a system which at best, will get you close to answering this.
However if you’re a fast-growth company looking to make the the next best decision to that, on how your marketing performance stacks up, then a combination of in-channel optimisation to get the best from your macro metrics such as cost per click (CPC) and clickthrough rate (CTR) on top of a blended view of marketing performance might just get your growth machine whirring that much faster.