Taking Attribution Seriously Requires More than a Tool

Shelley Eleby is a new contributor at The Makegood. Shelley is Vice President Marketing Services at quadrantONE, a joint venture of the media companies Tribune Company, Gannett Co., Inc., Hearst Corporation and The New York Times Company.

We’ve gotten used to the A Word: Attribution. In fact, since I was on the client side, we’ve come a long way from when we had no clue its importance. Over seven years ago we were practically fantasizing about cross platform media optimization/measurement and attribution and now these are our essential nomenclature. Now, no matter what side of the table, we buy into it. Understanding attribution is mandatory for performance and brand marketers alike. And, then, there are the essential tools: Nielsen, comScore, and all the usual suspects have come out with their own products for advertisers to run, command and measure relative performance across their own mix.

Yet, it seems we distract ourselves as an industry by watching the landscape on tool options and checking current wisdom on state of the industry – before we move a muscle. We can read “how to choose a tool” pieces all week long. Constantly evaluating tools can stall us in our tracks. You see, when it comes to getting serious about attribution, I am not certain this should be an industry driven strategy versus a marketer driven one. Marketers need to be pro-active in this area.

We must realize that our job is not done when we have picked our tools. Regardless of the tool, we must do the work to determine which model and approach are best for our own objectives and bottom-line goals. Of course, this happens in the context of the industry – but we must drive our choices from the inside.

Which Model Begets What?

Your model will vary based on what you aim to learn, understand or accomplish. If you are somewhat complacent, you may heed only the last click – crediting performance to the placement or asset that ultimately was a stepping stone to the conversion. This is an idle if not lazy approach, not taking into account the numerous media exposures and creative influences that came before. I like to think the industry has moved beyond this. Don’t be that guy.

Or, perhaps it is interesting to you to monitor that first point of engagement – the first click. This may be minimally valuable if you are studying purchase cycles and the continuum of engagement. But, I’m guessing your objectives are more serious. So, the first click, just like the last click – is basically inconsequential. Why go there?

Another potential model – one that at least shows a cognizance of all channels in your mix – is the equal weighting attribution model. While understanding relative performance does require you to show some discretion across the mix – if you use this model, at least you are not penalizing certain combinations while unduly crediting and rewarding others. It’s a neutral position, hopefully setting the stage for your doing something more thoughtful.

This brings us to custom models. This is where I believe we marketers serious about our objectives and bottom-line goals, must go. If we have a handle on what we want to learn, understand or accomplish – we should measure relative performance accordingly. Period. Get the tools, get situated and get going.

Because most of us – both brand and performance marketers – in the end care about efficient sales, maximizing output or ROI is our primary reason for being there in the first place. We’ll want to look at channel and placement level performance in a very granular way and build an algorithmic model that allows us to zoom in and do so. When looking at message, placement, creative, interactive mechanics, bids – we’ll want to examine touch points, interactions/engagements, and the relative performance of all attributes.

Operating in this way does require the tools set. But, as you can see, it has implications for talent, training, workflow processes – and a culture of getting it. Everyone involved – from strategist, to planner and buyer, to account management and analytics, must buy into the business objectives and embrace (day in and day out) the quest for cross-media clarity, performance understanding, attribution and optimization. You must put your team in a position to gain meaningful relative performance assessments. Furthermore, as you choose your media partners, be sure to share your attribution strategies and approaches. This transparency will position your partners to perform to the best of their ability, with your attribution model in mind. Only then can you intelligently benefit from and scale your own success.

  • aquarianalien

    Great article!