ADOTAS EXCLUSIVE — Despite all the healthy growth the online advertising sector has been experiencing over the past few years, it might surprise some industry insiders that there is still a large number of business managers (especially CFOs) who are not convinced about the value and accountability of online advertising. Instead of responding with the usual “they simply don’t get it” or some ROI-related cliché I am going to explore some reasons why online still does not have the credibility we all believe it deserves and what we can do to fix it.
The problem lies within
Several client-side marketers (including some Internet pure-plays) have made comments to me recently that “online advertising does not have credibility in my organization” or “my boss does not believe my campaign reports.” On digging a little deeper it became clear that most of the credibility problems boil down to metrics and the impact online advertising claims to have on the business. To be more precise, the numbers simply aren’t adding up in many cases.
It often comes down to a problem of perspective: on a channel-by-channel basis (premium display, search, network, affiliate, email, etc.) the people who are responsible for managing programs in those channels believe that they are tracking and reporting correctly, but when you add up all the numbers and compare it with what is happening in the client’s ecommerce or finance system, the sum of the parts is sometimes greater than the whole. And because most enterprise marketing organizations are still very much siloed – different teams, different technologies, different agencies – it is often very hard to tie all the data together to get a clear sense of what is really going on. But without such a macro or big-picture view, the component pieces will always suffer a crisis of credibility.
So instead of being defensive about the effectiveness of online advertising, let’s consider what needs to be changed for credibility to be restored.
Getting the big picture
In order to evaluate properly the impact of all your online advertising programs collectively, you need to create and implement a clear and transparent Media Measurement Framework which is part methodology, part best practice, and part technology solution. In simple terms, you need a set of rules and standards for measuring online advertising impact across your entire business that is consistently applied and justifiable to your managers. Here are some of the key attributes of a good Media Measurement Framework:
1 – Track everything
It is vital that all channels and all programs are tracked as comprehensively as possible. We often see situations where some channels are tracked but others (often email) are entirely ignored. We also see situations where delivery and response data is collected for all channels but conversion data is tracked only for a few channels. It is vital that you track everything on systems you control (to verify media owner data) and that you own the data kicked out by these systems.
2 – Integrate the data
Since there is no one system today that provides campaign delivery, optimization, and reporting for all online channels, we are all forced to work with multiple 3rd party vendors. There are increasing native integration points (especially for search and display data), but to get a fully integrated view of search, display, network, affiliate, email, and Web site data requires an independent data mart and reporting solution.
3 – Compare apples with apples
Use the same success metrics and counting methodology for all channels. For example, don’t evaluate one channel according to click-through conversions only and another according to click-through and view-through conversions. Or don’t track conversions in one channel using 30 day cookies and in another using 24 hr cookies.
4 – De-duplicate conversions
For the most part, campaign tracking and reporting is an internal evaluation tool, however, if you do any performance-based media deals (CPA deals, rev share deals etc), it is vital that you de-duplicate conversions from your various channels. Not doing so, and paying multiple performance media vendors on their numbers could mean that you are paying several times for the same conversion. You need a conditional filtering system (like DoubleClick’s Floodlight tag tool) to ensure that you only ping performance media tracking tags when they were the ones sending you the customer. And besides the duplication of conversions, it is highly reckless to implement media owner tracking tags directly on your order confirmation page as this gives your media partners full insight into your total sales volume.
5 – Make your metrics relevant to your business
Don’t simply track vanilla metrics such as “number of orders”. In most businesses the value of the order, the product(s) purchased and some additional industry-specific metrics (e.g. check-in date, check-out date, and room nights for hotel businesses) are vital in evaluating the real impact of a campaign. Using these rich metrics and combining it with additional offline data such as cancellations, returns, up-sells etc will enable you to evaluate your online advertising programs according to relative contribution (i.e. full P&L), as opposed to bland “cost per sale” metrics.
So what about “engagement”?
This discussion will not be complete without a discussion of engagement measurement. The term “engagement” has become one of the most widely used terms in online marketing and advertising in recent months and is being used in multiple different contexts (advertising, website usage, brand marketing etc). Like ROI and CRM, it is at risk of becoming almost meaningless.
However, for purpose of this discussion let me confine engagement to the way it is being used by Microsoft’s Engagement Mapping (EMAP) approach, i.e. a method of looking at the impact of all ad exposures and what contribution was made to the final sale, not just attributing a conversion to the last click. While a useful line of analysis, there are significant problems with the EMAP approach: First, it allows advertisers/agencies to make discretionary assumptions about the expected impact of particular attributes (ad size, creative type etc) in the absence of any statistical evidence that their assumptions are valid. This is like giving a toddler a loaded gun. If you can prove, by doing statistically sound data mining, that a specific ad placement contributed to conversions, then go ahead and attribute some of the credit to that ad placement, but don’t leave it up to the (inevitably biased) discretion of agencies/marketers. Second, to do engagement mapping properly one has to collect a massive amount of data and one has to integrate all channels. Doing engagement mapping with only a few preceding exposures and without including all your channels will simply give you a different partial view, as opposed to the real big picture.
Conclusion: Winning back credibility
The road to a robust and rigorous Media Measurement Framework can be long and hard, but the data is available and the technology to execute this approach already exists. The best approach is to create a roadmap that builds towards your ultimate vision over time and to get support for the project from multiple internal stakeholders. Ultimately, you will need to make some hard decisions, debunk some myths and take a stand. It is the only way you will ever gain credibility for your team and for online as a marketing channel.