How Stagwell’s SmartAssets is using data to bridge media and creative
The Media Leader Interview
Lindsay Hong, co-founder and CEO of the SaaS company, explains how big data can now be used to predict which creative assets work best in different media environments.
“Not all content is created equal.”
Perhaps that is an obvious observation but, for Lindsay Hong, CEO and co-founder of Stagwell’s “creative effectiveness platform” SmartAssets, the simple fact that creative decisions don’t often efficiently link up with media decisions presents an opportunity to correct a market failure.
Launched in 2023, SmartAssets is primarily a software-as-a-service (SaaS) platform with additional managed services. It uses machine learning to tag creative components from ads and then layer on performance data from those ads to understand what creative is optimal in a given media environment.
“What SmartAssets does is it creates a creative dataset that we’ve never had in marketing before and applies a whole bunch of scores, plus media data to it, to start to get some objective numbers behind what specifically it is about an image or a video that drives engagement,” Hong explains to The Media Leader.
The company began as a “passion project” out of multilingual content and media consultancy Locaria, part of Stagwell’s Brand Performance Network where Hong was chief strategy officer. The effort won Stagwell’s internal 2022 “Shark Tank” competition, beating 27 other applicants to win $1m in seed funding from the holding group.
Hong attributed the win to SmartAssets’ capacity to bridge creative and media insights through data to drive improved outcomes for clients.
“Creatives think that media people are like used car salesmen and media people think that creatives are overpriced haircuts,” she jokes. “That’s reducing it simply, but the client is stuck in the middle between all of these egos, all these people they’re paying money to.
“It’s emotional, it’s personal, it’s subjective, it’s difficult. We’re just trying to make that easier and more effective.”
Simple and specific asset optimisations
According to Hong, SmartAssets applies the same core principles as performance marketing — using data to make predictions about audience receptivity to ads — and applies them to creative decision-making. It does so for agencies and clients both inside and outside Stagwell.
Core services include tagging, analytics and optimisation, with the overarching goal being to create “effective content at scale”.
Tagging is centred on, in effect, transcribing creative assets by marking their every detail so they can later be analysed. Data points include colours, objects, whether the image is indoors or outdoors, who or what is depicted, whether brand titles and logos are visible, what copy is included, visibility of terms and conditions, and other regulatory content.
“We can pick up different product lines, we can pick up what’s in audio, whether it’s subtitled, what it’s saying,” Hong says. “You’re basically atomising the asset down to what’s in it at a very granular level.”
From that data — “as much as we can get, a minimum of a year and ideally three years’ worth of assets and performance data”, Hong outlines — SmartAssets offers a variety of optimisation possibilities.
Its SaaS Asset Governance Platform, for example, effectively works as a way for brands to automate quality assurance of their ads by applying brand and media best practice.
According to Hong, it is shockingly common for assets to be delivered from creative agencies to their media counterparts that don’t properly align for the channel in question.
“One of the agencies [at Stagwell] told me that 40% of the creative assets they get from another creative agency outside the network are the wrong-sized spoke for the hole that they’re intended in the media plan,” she says. “They’re not in the right dimension, literally the wrong shape.”
This most commonly occurs, Hong indicates, with TV ads that have been poorly repurposed for YouTube or static landscape images ineffectively produced into a square format.
“Just fixing that, making sure the ad is the right dimensions and resolution for the intended placement, already increases your effectiveness,” she points out. “That’s not very sexy, it’s not the most advanced analytics, but it’s pretty helpful.”
Informed by data science
The “sexy stuff” comes in via additional data science, such as measures around emotion and attention that SmartAssets can also check assets for.
Through its Predictive Creative Audit managed service product, the company can provide a list of the top creative decisions that have influenced whatever KPI it is a brand is looking to push. Creatives can then feed that information back into their process to, for example, make sure every asset for a given media channel has that specific asset included (or removed).
“The result is very specific creative recommendations for how to improve their performance, all based on your historical assets that your audience has been looking at,” Hong explains.
“Often, the client already has a gut feel on what has worked and what hasn’t, but they didn’t have any data that backs it up. [We’re] not coming in and saying creatives are wrong — it’s hardening those arguments, often celebrating really good creative decisions, as well as offering opportunities for improvement.”
She clarifies that SmartAssets is not in the business of taste-making, but rather bringing audience-specific media preferences to creatives so they can understand what works best for a given environment — in aggregate.
“The thing about media data is it’s at the sharp end of your customer engaging with your brand without really thinking too hard about it,” she argues. “A millisecond emotional response that can be aggregated into statistically significant insights.”
According to Hong, conducting such predictive A/B testing resulted in tripled return of advertising spend for an FMCG client.
Often, the recommendations for creative adjustments are simple, such as marginally reducing the number of products shown in a given ad or ensuring the name of the product is displayed prominently.
Alternative to DCO?
The process is also more efficient than many brands are currently deploying, Hong argues.
“If you look at existing creative testing solutions for media, like dynamic creative optimisation, the way it works is you create, say, 10 versions of an asset, algorithmically throw them up on the internet and you see which one sticks with a specific target audience,” she says. “You might say number four out of 10 is the best one, so you’ll just keep serving number four.
“But what happens to the other nine? They’re wasted. And you’ve already spent media on them, which has also been wasted.”
SmartAssets, she adds, is more interested in optimising creative before the campaign is live based on a brand’s historical asset data. The result is less waste and less guesswork.
Currently in development at SmartAssets are two “emerging services”, referred to as Influencer Intelligence and OOH Attention Reporting.
The latter provides predictive visual attention heat maps and descriptive metrics for OOH creative. Using a coded algorithm aimed at mimicking visual attention, SmartAssets can create a predictive map of where people are most likely to have their eyes drawn within a given OOH environment.
Influencer Intelligence, on the other hand, applies the same data efforts to brands working with creators on platforms.
Notably, clients have also asked SmartAssets to deliver insights on what types of creative assets work not just for their own brand on social, but also for competitors, since the creative data is all publicly accessible.
This is “absolute gold” for those wanting to understand how to get a leg up on their rivals on social, Hong says.
While Influencer Intelligence is relatively limited in terms of outcome data since the platforms are measurement black boxes, Hong does not think it likely that platforms would be open to partnering SmartAssets in the future to more easily access additional effectiveness data.
She warns that “what drives the platforms’ business is generally putting more assets up and spending more against them, which is not entirely what we’re aiming for”.
Hong notes: “What I encourage clients to do is to create the right asset rather than create loads and loads of assets.”
Havas Play Network’s CEO is on a mission to bring media and creative back together
This article has been updated after publication. A previous version of this article said SmartAssets uses eye tracking data for its OOH Attention Reporting product. The product does not use eye tracking data, but rather a predictive attention algorithm.