TMW #040 | Emotional analytics, Roblox the platform creator and email in the year of chaos

May 16, 2021

Welcome to The Martech Weekly, where every week I review some of the most interesting ideas, research, and latest news. I try to look to where the industry is going and make sense of it all.

👋 Why you should sign up

There are two versions of TMW, the full version for subscribers and a 50% version that's for everyone. You are reading the 50% version. Sign up to get the full version delivered every Sunday for this and every TMW. Learn more here.


Hello and welcome to new subscribers from New Zealand, Australia, The United Kingdom, The Netherlands, Brazil, and The United States!

Welcome to a milestone newsletter, #040!

This week I look at emotional analytics, the $19 Billion industry you haven't heard about which is built around collecting data on customer emotions visually, audibly, and on the web.

I also dive into Roblox, the platform maker that some call a next-generation media company, and how the pandemic changed the global state of email marketing.


Before we jump into things, I have two quick shoutouts:

  • I recently had a conversation on the Made by Marketers For Marketers podcast hosted by the talented Billy Loizou. We talk about the creator economy, my personal journey building TMW, managing fandom, iteration in media business, and defining the word “Martech” Link.
  • FS Martech, based in the UK, recently featured TMW’s latest podcast - have we hit peak Martech? The group, headed up by Richard Nolan is an online community that aims to support the use of marketing tech in the financial services sector. Pretty cool!

✍ Commentary

🎭 Emotional analytics. It always fascinates me that marketers are endlessly curious about how their customers behave. Maybe it has to do with sincere empathy, or a sinister will to compete. Either way, the Financial Times this week reported that the technology industry which collects data on customers' emotional states has ballooned into a $19 billion industry. As marketers everywhere have been collecting more first-party data, the silent but influential category of emotional analytics has gained significant ground not just in marketing but in security, HR, and other practices. Some of the world’s largest brands employ a level of emotional measurement on their customers, future employees, or staff, like Disney, BMW, Coca-Cola, and Intel, while some of the visual technology has started entering into public spaces like London’s Piccadilly Circus to analyze public emotional reactions to billboards. This technology niche is split into two major and one minor form of technologies; facial recognition, natural language processing, and behavioral interactions on the web. In just five years the size of the emotion detection market is about 3 times larger than email marketing and email marketing has been around for a long long time.

Now what’s interesting here is the way this data is collected. Traditionally marketers would run this kind of emotional analysis to make decisions on products, branding, and messaging through qualitative studies - get a bunch of people in a room to react to things. This is helpful to an extent but does not give you the ability to understand your existing customers. Now, new technologies allow companies to collect, analyze and attribute this data to specific customers at scale. This is done through the advances in machine learning algorithms and AI vision technology. One simple use case would be lead scoring. Say, a sales agent takes a zoom call with a prospect, the call records the prospect’s facial reactions to determine things like enthusiasm or fear, this would then naturally plug into the CRM system as just another data point. However, as simple as it might sound, the science behind measuring someone’s facial expressions is incredibly complex, making it the most nascent emotional detection technology at the moment.

This is because of the ability to measure the responses in people’s faces. How can an algorithm tell when someone is happy or sad based on a facial expression alone? There are about 70 - 80 different human emotions labels out there, with major cultural and contextual differences between each. Not to mention the inherent biases that make up the ML models anyway, and as most researchers have found the premise in which existing algorithms are developed lacks the kind of scientific research needed to ensure that the models don’t make discriminatory conclusions. The field which has more advances is in natural language processing space which, when contextualized to a person’s country, can be more effective by measuring someone’s intonation and language to determine a feeling. This is in a world with hundreds of millions of Echos and Google Nests out there along with the rise of audio social media like Clubhouse that is collecting audio data at staggering volumes. The field that is already commercialized well is using AI to measure behavior within online interactions within an app or website. Measuring something like a rage click is fairly easy to do and is a strong indicator of a customer’s anger, which can help marketers infer opportunities to improve experiences. As the technology is still in an early stage, it’s at a tipping point of real-life commercial applications and these vendors are betting big on the opportunity.

You might be wondering at this point, what’s the value of all this data? And I tend to agree with that question. A lot of this seems to be solutions looking for problems. And while there are some clear use cases for determining someone’s emotional state, the ability to use that data in real-time to improve an experience or target people with ads, for most brands, would be nearly impossible given the amount of data aggregation, streaming, and decisioning that would need to happen. The other side of the coin here is privacy and regulation, which strong European advocacy suggests that collecting intimate data like a facial expression or voice recording is a major violation of privacy without clear and express agreements to the terms of service. Regulation agencies in the UK and EU have recently stated that the ability to keep your emotional reactions private is a human right and will become enacted in policy soon.

And as we’ve learned from Apple ATT opt-ins, customers seem to be not cool with giving away more data than they need to. What’s interesting here though, is that the lines are drawn with machines. Humans naturally analyze the emotional states of others all the time, which is no violation of privacy, but when a machine does it with millions of people it becomes a problem and that problem is scalability. At this stage, emotional analytics is a fast-growing industry that is yet to realize promising competitive advantages for some categories. But for most, the data is not trustworthy or usable enough and makes people a little uncomfortable. Maybe we should leave the emotional analysis to our psychologists? Links: FT on emotion detection ($), MarketandMarket research, CallMiner, voice profiling, three futures for social audio, Wired on tech creepiness,

🎮 Roblox and platform creation. This week some very interesting stats came out indicating that more than 50% of US teenagers are using a gaming platform called Roblox. That set off alarm bells for me, namely because A, I don’t really know what Roblox is and B, that’s a significant population of the next generation jumping onto a game that allows you to create your own games within it. You may have heard of the concept of the metaverse here on TMW or elsewhere, and Roblox is a highly influential company in creating it. The premise of the game is that you can create your own worlds for others to play. Kind of like Minecraft, but the big difference here is that the games people create can be monetized, turning young twenty-somethings into millionaires in a relatively short period of time. The platform has around 8 million developers, with 300 billion hours of playtime. That’s some serious scale. And this week Roblox also changed its brand positioning from being a “game” to an “experience”. This was in response to Apple’s ongoing dispute with Epic Games, embracing their future as a platform creator within a platform itself.

This is a marketing dilemma because Roblox is a closed system that makes it very hard for brands to get their messages in there. And because monetization happens through creation within a locked social network, Roblox has very little incentive to insert ads within their platform. Despite this, Digiday calls Roblox one of the most important media companies of the future because of the sheer daily engagement metrics powered by the virtual currency that enables people to buy games within the platform. There’s already some interesting partnership happening with Hasbro and Monopoly to place products within the games people are creating, which is an early sign of future advertising avenues. We also have to remember that right now, the majority of US teenagers are having their formative years influenced by massive platforms like Roblox, building skills within the platform instead of learning to code. Roblox here seems like it's paving the way for new kinds of platforms on the internet, which places a premium on creation and creator while locking most brands out of participation. Links: Roblox: Game vs experience, Teenage use stats, platform usage, Roblox as media business, child safety on platforms.

📈Chart Of The Week  

📩 Email in a year of chaos. Acoustic have released a special benchmark report tracking the performance of email marketing in their platform during a year that was unpredictable for the channel. Sign up here to get a link to the full version and the chart. You won't be disappointed.


📰 Latest Developments

Google app transparency. In a similar move to Apple, Google has announced that developers will soon have to disclose to Android users what data they collect and how it is used. It will live in the Google play store in a dedicated safety section. The policy will also include an element of data security, calling out developers who don’t encrypt data shared in the app and the apps that don’t have adequate data deletion policies. Link

eBay launches NFT market. eBay are known for bringing the auction to the internet, and now, in what feels like a big step change, they are now opening the door to crypto payments on the platform with NFTs first and then soon with product purchases. This is promising for eBay which has certainly lost the marketplace war between Facebook and Amazon. This might give them a new angle in the market. Link

Twitter planning subscription service. At the time of writing this, the idea of Twitter launching a subscription service is highly speculative. However, the features do look interesting - $2.99 a month could give you an ability to collect favourite tweets, super follow others and the ability to undo posts. Link

📚 Reading

Starbucks, the tech company. It’s no secret that Starbucks got into the digital game earlier than almost every other hospitality business. But now it’s more of a tech company than a coffee brewer. The company boasts 25 million mobile payment users (more than Google), has 23 million active digital rewards members and rivals many banks with the amount of capital housed in the platform with the Starbucks Card. No other company has been able to command the kind of loyalty in customers that Starbucks enjoys. Link ($)

Seth Godin on the blockchain. This was a surprising read to me, however Seth Godin is thinking about how the blockchain will impact marketers in ways that many haven’t. His point is that most of the products we use online today require databases, the blockchain is just another database. But because it’s an open, distributed database it will fundamentally change how technology is applied to markets. Like the printing press, or the web today, most people won’t care about the tech but will pay to enjoy its benefits. Link

What’s next for the cookie? With all this talk of the shutdown of third party cookies, you’d think that the whole technology of web based cookies is doomed. Good news is that it’s not. There are still dozens of cookie solutions out there including Unified ID 2.0, which is a framework that allows for cross site targeting and attribution but with far greater controls around how it’s used. Link. Also see my colleague’s very interesting deep dive into HttpOnly cookies as a viable alternative solution. Link


🔢 Data & Insights

Cloud services and enterprise data. New research from MIT and Databricks looks at the trends in how data is managed in enterprise companies with more than $1B in annual revenue. More than 50% of CIO/CDOs surveyed said they are struggling to scale ML use cases and only 13% of orgs are delivering on their data strategy. Link

Amazon’s ad supported video. Prime streaming is one of the only subscription services that also include paid, targeted advertising as pre-rolls before shows. Amazon has released numbers on effectiveness and reach with a recent IAB event. About 120 million people are reached per month. Amazon is an advertising business. Link

Environmental impacts of the blockchain. The Cambridge centre of alternative finance tracks the near-real time impact on energy consumption as Bitcoin is mined and purchased around the globe. The upper bound is 150 TWH, about the amount of energy needed to power Egypt or Malaysia. However, clean energy like existing hydro plants could power Bitcoin more than 28 times over. Link


💡 Ideas

The changing state of the spreadsheet. Excel hasn’t changed much at all for two decades, then Google Sheets came along and put it into the cloud, and companies like Notion and Airtable repositioned spreadsheets into other workflows. This new product, Rows takes an API to the spreadsheet. It’s built to have seamless integration with popular apps to pull, manipulate, analyse and push data. Link

Failure Theory. Ever had an engineer “accidentally” delete the golden record in your company’s database? You’ve probably heard of horror stories like this, as have I. Here’s an interesting take on the idea of catastrophic failure and complex systems. The hypothesis is that the more complex the system, the higher the risk that there could be a highly consequential failure as these systems tend to optimise on preventing smaller failures that are more likely in probability. Link

Real-time interaction management (RTIM) is an emerging concept that centres on the “last mile” delivery of customer experiences based on in-the-moment decisioning. It’s garnered enough attention to have a Forrester Wave commissioned for it. The idea is that customers can have needs that change on the minute and being able to orchestrate responses across channels and at scale seems to be a need worth looking at. Link


✨ Weird and Wonderful

Alpaca moderation. The Chinese version of TikTok has started blocking posts that feature alpacas due to the rise of highly popular videos that stream grooming and dressing up the animal. The decision to block however did not come from human moderators but the artificially intelligent content screening system. Should we be a-llama-ed? Sorry. Link

Lovetris, Tetris for millennials. Tetris is usually pretty hard, this version uses AI to only give you blocks that will definitely fit every time as you play the game. That way everyone can win, and no feelings are hurt. Link

WeWatch. An app that’s like Tinder for movie preferences. You and your partner download the app, swipe right on the movies you like and it will tell you if you have a match. This app could save so many lounge room fights. Link

💼 Come work with me!


A lot of what's featured here in TMW is the product of the rich conversations I have every week with the talented people at The Lumery.

We're going through a growth curve, and we have heaps of new job openings going across data analytics, marketing automation, business analyst and customer strategy.

🦘If you're an Aussie and love working in marketing tech then take a lookhere.

Reply to this email if you'd like to ask me any questions.


Stay Curious,

Make sense of marketing technology.

Sign up now to get the full version of TMW delivered to your inbox every Sunday afternoon plus an invite to the slack community.

Want to share something interesting or be featured in The Martech Weekly? Drop me a line at juan@themartechweekly.com.

Juan Mendoza

Marketing technology strategist at The Lumery, I analyse marketing, data, and technology trends for some of the most well-known Australian and global brands.

Great! You've successfully subscribed.
Great! Next, complete checkout for full access.
Welcome back! You've successfully signed in.
Success! Your account is fully activated, you now have access to all content.