Whilst gaining knowledge of your users’ cross-channel and cross-device behaviour along the whole breadth of their interaction journey, being able to generate growth, cost-savings and better allocate budget is the real key to success.
This is exactly what helped Equifax achieve a $1.25 million media spend saving back in 2015, and more recently a reported 61% increase in ROAS for Bupa in 2018.
With a small amount of legwork, you too will have the power to create positive change for your organisation through marketing attribution strategy. The first step is to get your ducks in a row, and perhaps most importantly, be able to bring your online and offline activity together for a complete picture of the customer journey.
The whole picture
If you’re a business that doesn’t solely operate online, then by looking at marketing attribution for just your online activity will be limiting. It’s certainly better than giving no consideration at all, but you’ll receive no clarity on how your website visitors or app users engage with your stores, locations or sales representatives, including where the touch points are, and how these all influence the overall user journey.
By giving users unique IDs that can be recognised cross-platform, you can tie significant activity such as purchases or sign-ups to them via your CRM system. When a user enters a physical space to purchase an item, you can also record this against their unique ID in many ways – most commonly by implementing a loyalty card program or receipt-via-email service.
By shining a light on the proportion of users that perform their research online and proceed to purchase offline, further value can be presented and attributed to your marketing channels accordingly, valuable data that would previously be invisible.
Painting the picture
Once you have all your data and sources identified, you’ll also need to present this data in a digestible way for all parties involved. The way to best do this is to integrate your CRM platform with your chosen analytics tool and/or data visualisation platform. Arguably, the best data visualisation platforms on the market at present are Tableau and Microsoft Power BI.
The biggest blocker
The real challenge, of course, is getting the buy-in of your company’s senior stakeholders, which is why it’s vital to present the data in the most understandable manner. If their focus is on bottom-line figures, why would they care about attributing value cross-channel?
Present your case in the simplest terms: the reduction of wasted marketing spends in areas that aren’t pulling their weight, coupled with enabling a marketing strategy that is both effective at driving growth, in-line with today’s online habits.
Keeping things easy to understand will almost undoubtedly help your cause, as it is easy to get caught up in the particulars.
Planning your integration
The first big move in the right direction is to integrate your marketing platforms, along with your CRM system, with your analytics/visualisation platform (See figure 1). By doing so, you can tie all relevant activity both online and offline, particularly if your business relies on store visits or purchases, or call centre activity, for example.
- Microsoft Dynamics presents a handy advantage in that it has a straight pipeline into Power BI, both being Microsoft products. By combining the two, you can make a start on your attribution system right away.
- For Salesforce, you can make the process easier by plugging directly into the Google 360 suite, should you be paying the usage fee required to use the platform (at time of writing, upwards of $150,000 USD.)
- If direct linking is not an available option, for instance if your platforms are not cross-compatible or if using the standard Google Analytics platform, then you will need to enable the ability to create a shared unique identifier (such as the unique user ID described earlier on) to tie user activity across platforms.
If you have multiple CRM systems or data sources that need to be integrated, this can prove to be difficult. An alternative to linking all of your systems is to instead opt for data warehousing. By combining all of your data into one local, linkable source, you eliminate the reliability issues that can come with linking many sources (and relying on their own APIs to function simultaneously) (See figure 2).
Google’s Big Query can allow you to achieve this, alongside other systems such as Amazon Web Service’s own data compiler tools. Once all of the data sources have been combined into one centralised “master” list, you can perform one single integration into your analytics/visualisation platform.
Observe, then grow
Once all is properly integrated, calibrated and visualised, you can then begin to identify the areas with the biggest potential to drive growth with some more budget and/or attention, the areas of “hidden” benefit (See figure 3).
On the default last-click model, it would appear that Facebook is the main driver of purchases with a below-average CPA, certainly more favourable when compared to YouTube. Therefore, it would make sense that Facebook has been given the majority of the budget, and that YouTube has been running on a much smaller share that heavily limits potential ad exposure.
Moving to a multi-touch model reveals that YouTube has been greatly assisting the journey to conversion, with a significant portion of users first seeing an ad on YouTube before purchasing later after clicking an ad on Facebook.
By increasing ad exposure on YouTube (by adding budget) to attract more users earlier in the journey, the next period’s results then appear as seen in figure 4.
By boosting the first touch point, the overall result is much more favourable with a 44% increase in return for just a 9% increase in ad spend.
Had the last-click model been kept, it would’ve seemed illogical to increase budget and in doing so would have shown an inflated CPA figure. Therefore, the overall benefits would never be realised or attained.
To maximise effect, you don’t want to stop at the over-arching attribution model, looking at all of your marketing channels as a whole. Consider applying a non-last click model to each individual platform as well (where available.)
For instance, you may see that your Google Search Network activity may be great as a first touch-point and/or a large contributor to driving the conversion as a last touch, but this doesn’t tell you which campaigns, keywords or ads are working best within each “role” (See figure 5). This will aid the process where possible, and certainly where multiple touch points are experienced within the same traffic source.
Compare, test & optimise
The realisation to take away is that no attribution model is ever really “complete.” Over time, customer behaviour will change, as will the business. Attribution needs to be a “living” concept that is continuously adapted to most accurately reflect value and efficiency.
It’s important to test new ideas and compare them to what you have established. You may find new pocket areas of hidden benefit that will allow you to re-optimise your optimisation processes and generate even more business benefit.
Ready to start
When you consider the amount of media spend and work effort that goes into your marketing, alongside the continuous development of your website or apps, it makes perfect sense to put data at the heart of your business and take digital marketing attribution seriously. Is your business ready?