AI, Attribution Models, and Luxury
FSW explores how AI-powered attribution models can help luxury brands measure content strategy efficacy.
As every consumer-facing organization is effectively omnichannel, content strategy is an essential toolkit to ensure that you are reaching your audience in the right way and doing so consistently over time and touch point. Content strategy can serve as a glue between your broader corporate strategy and your brand story and between current and future decision-makers. If you happen to work in a luxury brand then content strategy will need to take on unique characteristics not found in other retail sectors owing to luxury's need to convey a sense of rarity, scarcity, and even magic that draws in current and future consumers.
Yet, how do you know if your content strategy is working?
It is straightforward to list all the scalars that go into the vector of a core content strategy, and there are a lot of places to find such lists, but measuring results is less straightforward. For luxury, measuring results is even trickier as a lot of brand storytelling is geared towards selling an image or feeling. Also, luxury storytelling can be more oriented towards the longer-term than other retail sectors by appealing to aspirational buyers.
In standard marketing, attribution models are the core workhorse to quantifying the return-on-investment (ROI) from campaigns. In this essay, we explore using attribution models for evaluating your content strategy with a heavy focus on the steeper climb that it takes to understand when content strategy is working for a luxury brand and when it doesn’t.
What can an attribution model offer you?
An attribution model is a way to assign credit for making a sale. Consumers can come across an advertisement or product reference in a wide variety of places in a firm’s marketing funnel: a social media post, google advert, direct email, website visit, a permanent store installation, a pop-up store installation, or these days even on a video game, Roblox instance, or NFT minting. An attribution model is supposed to provide an idea of which channel or combination of channels persuaded a consumer to buy or “convert.”
Some of these models are pretty simple and give all the credit for a sale to the first place that the customer encountered the product (first touch attribution), last place (last touch), or weight all the touch points equally (linear attribution).
Slightly more sophisticated models will overweight the first and last touchpoints and underweight touchpoints in between (positions-based) or assign the last touchpoint the highest weight and preceding touchpoints will receive sequentially lower weighted (time decay).
If data are sufficient, then econometric or machine learning models can be introduced. These models have the benefit of being better rigged to provide higher quality forecasts and can be updated at as high a frequency as you have sales and marketing data to plug into them. The difficulty can be reaching the threshold of obtaining enough quality data, that is reliable and valid, to make those models work.
Attribution models and content strategy
Though it was not their intended purpose, attribution models can provide insights on the effectiveness of content strategy.
A content strategy helps you ensure that your stories are aligned with your overall business goals. For example, if your business objective is to increase website traffic, your content strategy should focus on creating stories that are optimized for search engines and promote social sharing. After all, what is the utility of great stories if they do not reach your target audience (or your aspiring audiences) in the distribution channels where they are to be found, whether that is in a store, on a particular digital platform, on a corporate intranet site, in the metaverse, or on a video game.
A content strategy helps you measure the effectiveness of your brand stories and keeps strategies fresh and aligned with evolving corporate needs. By tracking metrics such as engagement rates, click-through rates, and conversion rates, you can determine what is working and what needs to be improved. This allows you to make data-driven decisions about your content strategy and ensure that it is achieving your business objectives. This results in a feedback loop of learning between your corporate strategy, your company stories, and your content strategy.
To make that feedback loop work you need to have an empirical basis of attribution. The features of a content strategy attribution model might overlap with standard marketing attribution models, yet would also have some unique features. Content strategy, like content marketing, is not just concerned with the locus of customer-brand messages that drive consumer engagement, but also the type of messaging that drives convergence. For example, among other things, content strategy is about getting the right message in the right tone to the right audience at the right touchpoint. Is it advisable to use the same type of content on an IRL versus digital channel or on a video game versus a metaverse-like experience? Is differentiation better?
Measuring which content is effective in which instances would be the heart of its attribution model. Some of the variables of interest, for instance, content tone, are more qualitative in nature. When we work with clients, we teach them how to code qualitative data into quantitative variables that can be inputted into attribution models.
What can a luxury brand get from a content strategy attribution model?
Luxury products have unique characteristics that require a unique type of content strategy. Luxury brand stories need to make consumers and potential consumers feel a certain way about their product. Luxury is a weird business in that, unlike in almost every other sector, higher prices may lead to increased demand by signaling quality. This works for several reasons that we understand somewhat intuitively. One is the prestige factor. Another is the experiential connection that luxury brands create with their market. Customers can value the experience and anticipation of shopping for luxury products.
The job of brand stories is to create that feeling and weave together commerce with a sort of magic. In fact that magic is does it feel special to engage with a luxury brand versus a non-luxury brand? Does it feel unique? Does it feel personal? Does it justify your time and treasure? In this sense, you might say that luxury is not about goods or services. It is Experience as a Service (EAAS).
Creating that type of engagement requires a content strategy that focuses on a carefully crafted brand story that communicates the magic across very different types of platforms. Communicating the magic in a flagship store is one thing, but can it also be done in a video game or digital immersive environment? The answer is “yes” but the requirements needed from the content strategy are that much more demanding.
In a previous post, FSW constructed a journey mapping that a Gucci customer might follow along with the Kering brand’s customer touchpoints. It is an example of good storytelling across media: from IRL to digital to the metaverse. What a good attribution content strategy model would give you is a way of assessing the value of each node of the journey and an assessment of which messages helped move the customer along and which may have pushed the consumer away.
Ultimately, this sort of attribution model is valuable across a luxury organization. This helps with brand storytelling and marketing, of course, but also supply chain management, pricing, and backend functions that assemble each touchpoint.
Whither AI?
The question that we receive from clients and readers all the time is can all this be automated somehow with AI? As noted, AI (particularly machine learning) certainly can play a role in generating insights from your attribution models provided you have enough data to the most out of those models. Machine learning is particularly adept at pulling together insights from large datasets and with a large number of predictor variables.
However, AI is no more a magic bullet for generating good attribution models than it is in generating text. To get the most out of marrying AI and attribution models, luxury or any other kind of brand needs to design its customer touchpoints so as to collect the right types of data. These data need to be reliable (so that results are stable across time and types of measurement) and valid (measuring what they are supposed to measure).
Content strategy when married with AI and attribution models can produce big insights when these conditions are met. It certainly requires a lot of mindful planning of brand storytelling, refined data measurement methods at each touchpoint, and well thought out content strategy. Luxury works because it sells magic. But how to create that magic is something that can be the function of strategic design.