Data, data everywhere, not a metric to relate to.

The real world is a messy place… marketing exists in the real world, unsurprisingly, this leads to messy information. We’ve discussed in other articles how to distil data into meaningful reports, the consideration that precedes this is how are you gathering that data.

Here, we are going to explore and discuss the various ways tracking works, and what its limitations are in digital marketing.

Determine a source

All reporting in the business should come from a central source, this article won’t be a discussion on CRMs vs DMPs vs web analytics vs advertising platforms (not this one, see why later). They all have their place, every client I’ve ever worked with has their own unique methods.

Systems are like departments, fit for the purpose they serve. As we’ve done in previous pieces, we’ll focus on the marketing department for top level reporting, however it is crucial that marketing work with Finance, BI (business intelligence) or MI (management information) teams for this first part.

The key is having one place for consistency, but also to ensure we don’t trip over our own reporting.

The key part here for Digital Marketer is to set out their stall of activity – what activity do we expect to influence bottom line reporting “cash in the till”. This is what the business wants to know typically day-to-day, week-to-week, etc.

Topline: how is the business performing and what aspects are driving it?

To do this, we must first define where this data is collected and how…

Those pesky ad platforms

Too frequently businesses allow their media teams (whether in house or external partners) to report verbatim from Google Ads or Meta Ads Manager (to name but two), however this has little consistency with real BI or MI data reported at higher levels.
The reason for this is owing to de-duplication, a process that seems to be a subject reserved for discussions in darkened rooms when it comes to reporting, however it is pivotal to increasing business efficiency and effectiveness.

In its simplest form, de-duplication is ensuring credit is given to a single entity.

For example, if you are running Google Ads and Meta Ads, it is entirely plausible for both platforms to say they drove 100% of your sales that week, great we have 200% of sales… hmm…

How?

Well, without going into Attribution management (see below), they both have their own tracking pixels and techniques that unsurprisingly don’t speak to one another Google cannot see Meta’s pixels and vice versa.

User clicks a Facebook ad -> comes to website likes what they see -> leaves
User searches on Google for your brand to find the product again -> purchases the product

Based on the typical window that Facebook holds this cookie data for 7days, their platform would suggest that the journey was;

User clicks a Facebook ad -> comes to website likes what they see -> purchases the product

Yet, we can see Google deserves some credit (maybe not depending on your view, a subject for another day).

Therefore, this is why when presented with an isolated Paid Social, Organic Search, Affiliates, Display or Paid Search report, check where the data is from…

Total business sales from BI/MI = 200

Paid Social (Meta data) = 80
Organic Search (Google Analytics) = 60
Paid Search (Google Ads Data) = 70
Paid Search (Bing Ads Data) = 20
Display (Ad Network Platform) = 15
Affiliates (AWIN) = 30

Marketing results = 80 + 60 + 70 + 20 + 15 + 30 = 275

Attribution, Cookie Windows and Data Collation

Furthermore to this, the platforms have over time progressed to change and ultimately create confusing models if they’re not understood.

Attribution is the method in which credit is granted to a channel, campaign, ad, etc.

Historically the digital industry reported on a purely last click model – all credit to the last touch point – this is down to technological simplicity;

Tracking a link click -> “watching that user on the site” -> product purchase

We’ll discuss more about attribution in future pieces.

However, we discuss a lot, the difference between reporting data and optimisation data and this is where it is crucial to understand this. Different attribution models can be helpful for optimisation data.

Optimisation data is a data point for the advertising operator to make a decision with, i.e. a Paid Social Manager determining where to move budgets or bids and when. Think of this as an indicator that a user has been engaged from an ad we have put out – how we determine that engagement can vary, a button click on the website may be enough to say to the operative “do more of that ad” rather than always a sale.

However, the board rarely wants to hear “X people have clicked a button” so the definition of a ‘conversion’ needs to be clear and centralised knowledge for Reporting purposes.

The same is true of attribution models and cookie windows, the operative can glean a lot from a different lens of the data. To rely solely on a last-click model can be a perpetually spiral campaign to lowest denominators.

Consistency in data and understanding ‘results’

We’re not advocating for Last Click to be the way to funnel your data into MI reporting, it’s generally the simplest, but the key is consistency.

Our advice is to stick to one path of collection and delivery, that easily aligns to the businesses world view. Therefore pushing marketing intelligence into your core internal reporting is crucial, in our view.

You can use a different model to power your MI reporting, if you wish, however it can have gestational issues and a dynamic nature that can be hard to explain outside of marketing teams.

Parting thoughts on Privacy

There’s a lot to be said about privacy in this regard, especially in the UK and EU following GDPR, so consider carefully how you are apportioning and appending data.

We are solely advocating a single way of reporting, and validation of reports supplied by the team buying ads on behalf of your business.

Furthermore, between Cookie Policies, Ad Blocking, private browsing and general privacy awareness, the tools of yesteryear are less and less effective – owing to their Cookie reliance. Therefore understanding where and how data is collected is more important than ever for marketing teams.