It's not how big your data is, it's how you use it

Simon Williams, National Head of Strategy and Product at Columbus
By Simon Williams, National Head of Strategy and Product at Columbus | 4 March 2015
 

Recently, Facebook launched a new way to measure conversion that implies that current online measurement is ‘outdated’, that the industry is ‘lagging behind’ and that clicks alone are ‘useless’. This latest measurement development takes causation into account when measuring conversion but I believe that this is fundamentally incorrect and that clicks are still vitally important to digital media.

The key to effective measurement and to unlocking potential is attribution.

Big data without good analytics can lead to bad decisions

To thrive in this industry, the data we collect has to be intelligent; it has to be qualitative, actionable and informative. Data is the life blood of our media mix and optimisation.

The article references great results achieved for Open Colleges which were collected over a 10 day period. However, I question whether 10 days is really enough to achieve data wisdom. Collecting historical data allows you to judge who your audience is despite seasonality, campaign length, changes in media spend and general brand sentiment.

In order to stand out in social media you have to fit in

We understand that Facebook is trying to connect meaningful advertising with the subconscious mind in order to influence behaviour, and therefore assume conversion uplift.

However, post-impression analysis and tracking isn’t as simple as claiming an eyeball on a newsfeed as an active engagement – it’s dependent on numerous factors such as the creative used, the campaign’s reach and frequency of the advert.

Attribution and post impression analysis is no small task indeed, so let’s not mention the challenges that arise in measuring and attributing multiple channels, keywords and layers of look back cookie windows. Imagine now, cuddling above the line advertising, the issues surrounding cross-device tracking and ROPO (research online – purchase offline) behaviours.

Attribution in real time

Attribution in real-time is essential when it comes to measuring the impact of brand when a final conversion is made online or in store.

It is therefore imperative that brands understand their target audience before initiating complex performance or brand campaigns. We have to be in the right place at the right time, with the right message.

With the help of real-time attribution models, we’re able to gather both campaign data and offline store data, allowing us to make wiser decisions.

But perhaps the biggest struggle within the different reporting systems and different methods of measurement is that all claim to be that ‘single source of truth’. The data analyst now has to determine when and how to value a display or social impression. Layer on custom look back windows, and hey presto – there is your digital attribution.

We wish!

Post-impression tracking may be one of the most complex methods used to understand what actions a group of individuals take after being exposed to a piece of advertising – but having not clicked.

So what do we suggest?

When starting to reallocate credit to channels based on their position in the customer’s journey to a conversion, it is important to thoroughly test and assess data to ensure there is an adequate sample size and external factors that may influence the results are accounted for.

Other Facebook offerings such as Audience Insights, targeting via Custom Audiences and Lookalike also provide an extremely innovative way for brands to understand and interact with potential consumers.

Channels like Facebook which are intrinsic to many consumers’ lives are getting on the front foot in terms of calling out the outdated nature of click-based measurement across digital. It makes sense that their "conversion lift" tool is opting to instead measure uplift in sales from ad impressions given their knowledge of how impactful the platform can actually be. Here in Australia with over 15 million users logging in each month, there are only a handful of other platforms that can guarantee that level of breadth.

Post impression tracking and attribution isn’t new, but it’s a proven method of accountability in online advertising.
Clients and the industry need to look to a data provider / agency that is channel neutral in their findings, leverage a data warehouse and/or media buying platform to execute, in order to make less ‘wrong’ decisions in today’s fragmented digital environment.

Ultimately, it’s not how big your data is, it’s how you use it.

By Simon Williams, national head of strategy and product at Columbus

comments powered by Disqus