The Futurist: Revisiting Attribution

A quaint houseboat floating on a sparkling Dal Lake beams from the magazine page and catches your eye. You’re conscious of the accompanied text, but it’s the powerful image that’s grabbed your attention.

You’re reminded of the effect the image had on you as you stumble across it yet again, at another time, as a website banner. However there’s a far greater impact as you take a little more notice of the caption which reads: “Up to 35% off houseboats in Srinagar.”

Game, set and match, you’ve decided your next holiday destination.

You explore further by clicking on the banner, and spend a good deal of time supplementing your research with Google. Overall you’re pleased to have selected a holiday destination that strongly appeals to you and suits your budget.

This is where the fun begins for the marketer and data analyst!

“Attribution” is a fairly dreaded buzzword among marketing professionals. You have all your company’s marketing channels integrated, synchronised and designed to suit, and the competition of which works best and for what purpose, is microscopically evaluated.

What’s producing the most site visits? Which channels results in revenue, ROI and how does it serve the KPIs?

In the context of the houseboat in Srinagar, what was responsible for the conversion? The offline ad, contextual banner or the paid search text ad?

Marketing in a digital world has effectively added to the complexities and fragmentations of reaching a consumer. Add to that mix user-generated content, independent brand messaging, the increasing importance of word of mouth, and you have an interesting and incisive analytics challenge.

Traditionally, most companies have used the last click approach, which is fairly straightforward, to understand, report on, analyse and act upon. Many brands still use it and while it has its obvious merits, the proliferation of channels in itself poses problems aplenty. Discounting offline as a means of measurement by online tools is still a topic of debate.

Learnings from what role each channel plays in the purchase funnel, and their likely impact post-purchase in terms of loyalty and reinforcement is a first step. Determining the most appropriate channel mix and incorporating the flexibility to refine this mix as required is the next. It’s not just a question of having the right tools and systems, but also the ability of changing mindsets, acknowledging that not all channels deliver the same way or serve the same purpose.

There are a few inherent problems with trying to model attribution.

  • Incremental value added by new customers compared to the value derived from building a loyal customer base.
  • The perennial offline versus online, and how to piece them together.
  • Flexible factors such as seasonality, product packaging and pricing to name a few.

Controlling all of them is an uphill battle, but at the same time, it doesn’t negate the necessity, or usefulness of attribution.

Attribution and modelling is an imperative part of most businesses today, particularly in the world of e-commerce. With the growth in app-related e-commerce activity, it’s vital to understand customer behavioural patterns from a variety of sources and devices, rather than generalise based on similar company experiences. Each is unique.

Testing multiple attribution methods simultaneously (last click, position-based, linear) is a good place to start to gain a better understanding of the impact of different touchpoints.

Once you’ve established a solid understanding, it’s a good time to then explore methods to customise the attribution model to suit your business. Cross device attribution through different means such as log-ins, fingerprinting, or good old extrapolation based on historic data, would tie in splendidly here.

Packaging it all together – attribution, devices, measurement, data – and then assigning suitable weight to a variety of channels and devices helps the marketer to better understand the value of each stage of the purchase path. Instead of just a bird’s-eye view of the entire customer base, it will facilitate a more accurate segmentation and identification of users. In addition, you’ll have the potential to unlock lifetime user values and understand their unique journey, which in turn, results in more effective marketing, resulting in a more longterm and sustainable relationship.

Ultimately, nothing wins out more than delivering an outstanding product or service at the best prices to suit your target market. Attribution and modelling will help seal the deal by enabling more effective communication via the appropriate channels.

Measurements and resulting adjustments to messaging will deliver a strong business performance, happy and satisfied customers and, even more importantly, brand loyalty.

The writer is Divya Ramaswamy, vice-president of digital marketing, Wego.

This article was first published in Marketing Magazine Singapore’s Jan-Feb 2016 print edition. To read more views from senior marketers click here.