75% of advertising is wasted, and who is going to dominate in identity wallets?
75% of advertising is wasted, and who is going to dominate in identity wallets?
Plus: Biometrics are not credentials, The Unwinnable Cost Problem of Identity Verification, and how Humane’s AI Pin works
Hi everyone 👋
Thanks for coming back to Customer Futures. Each week I unpack the disruptive shifts around Empowerment Tech. Digital wallets, Personal AI and digital customer relationships.
If you haven’t yet signed up, why not subscribe:
Music has just changed forever and we should be freaking out more about it.
This mini-intro isn’t about personal data. Or about privacy. Or about Empowerment Tech for that matter.
It’s about AI. And what just happened to music.
Quietly. Disruptively. You need to read the full piece, but here’s a snippet from the recent mind-blowing post by James O’Malley.
“A couple of weeks ago, a company called Suno released a new version of its AI-generated music app to the public. It works much like ChatGPT: You type in a prompt describing the song you’d like… and it creates it.
“The results are, in my view, absolutely astounding. So much so that I think it will be viewed by history as the end of one musical era, and the start of the next one. Just as The Bomb reshaped all of warfare, we’ve reached the point where AI is going to reshape all of music.
“What’s important is that this is just the beginning. All of these songs merely represent the worst this technology will ever be.
Why should we care?
Because some experts believe we’re living through multiple exponential waves of AI progress. That means several related things are going to combine and reinforce at the same time. Speech. Video. Compute.
Humans aren’t very good at understanding non-linear progress. Remember that ChatGPT was released less than 18 months ago. And now think of a part of your job that doesn’t look likely to be impacted by it in some way. Yes, digital content. Yes, software development. Yes, email and presentations and creative.
But think of task scheduling. Think HR. Think finance. Think innovation and strategy.
Now start thinking about that speed of change, and the impacts, across Empowerment Tech. Digital Wallets. Personal Vaults and Personal AI.
Our digital lives are about to fundamentally shift.
And when we won’t be able to trust anything we see or hear, we’re going to need new digital trust infrastructure. That can’t just be left to just the BigTech players like Apple, Google, Facebook and Microsoft.
We’re going to need a new breed of digital tools. New thinking around digital assurance and governance. And a new mindset when it comes to digital privacy, digital security and data.
Which is why I’m still writing this newsletter myself.
AI will help with topics. Will help me remember. But I believe that if AI is responsible for pumping this stuff out, then we lose trust.
We lose authenticity.
And it matters because I’m in the connections business. Connecting ideas. Connecting stories. Connecting businesses and value.
And connecting you.
Digital trust is at the core of all that.
And of course, it’s all about the future of being a digital customer. So welcome back to the Customer Futures newsletter.
In this week’s edition:
75% of advertising is wasted, but most is fraud, inefficiency and the wrong KPIs
Privacy policies should be required to include a "how we make money" section
The Unwinnable Cost Problem of Identity Verification
Patient data traded without us knowing - it can’t be solved, can it?
Who is going to dominate in identity wallets? It’s the economics, stupid
Biometrics are not credentials - and here’s what we can do about it
How does Humane’s Ai Pin work?
… plus other links about the future of digital customers you don’t want to miss
Let’s Go.
75% of advertising is wasted, but most is fraud, inefficiency and the wrong KPIs
Post of the week from the excellent Tom Goodwin. Pasting in full here for its awesomeness.
“1890 - “Half the money I spend on advertising is wasted; the trouble is I don't know which half" - John Wanamaker
2024 - "75% of the money I spend on advertising is wasted, but some of it is fraud, some is inefficiency, some of it is chasing the wrong KPI's, but that keep the CEO happy , and I could find out what's going on, but my power is derived from my media budget, and I'm scared to find out what's really happening and also short term metrics are far more important, as I’m forced to hit quarterly targets”
“I’m just saying there are incredible opportunities out there to do advertising and marketing in a much more compelling and efficient way, but somehow we’re a bit lost.
“I think the starting point is acknowledge how far from ideal things are, and to realize quite how different things could be if we were more ambitious about the way we use technology and more empathetic about what really works and matters.”
More ambition? Yes please. EmpowermentTech is well-placed to drive a stake through the heart of today’s digital ad ecosystem. Though Tom’s “we’re a bit lost” may be the understatement of the year.
It’s because the customer is nowhere to be seen in today’s digital advertising hellscape.
There’s no ‘digital customer’ to shake hands with. At best, businesses have an email address or phone number, a (guessed) digital profile to target, and a digital cookie with a half-life that degrades hourly.
With EmpowermentTech, the customer can finally show up as themselves. And it will be the customers themselves that reshape digital advertising as we know it. It’s going to come down to digital wallets (who), personal vaults (where) and personal agents (why and how).
Digital advertising is estimated to be ‘worth’ (erm) at least $700bn.
If you take 1% of that and point it to real people (solving fraud), with real data (solving quality) and understanding real context (solving conversion), then $7bn looks like a pretty good ‘customer engagement strategy’ ROI to me.
Privacy policies should be required to include a "how we make money" section
While we’re talking about digital advertising, in 2012, FB bought Instagram for $1bn. Insta now makes $1bn in ad revenue every week, and accounts for nearly a third of all Meta's revenue.
Why?
Because apparently, Insta is one of the best places to put digital ads online. But only when ‘best’ means fractionally less crappy than all the other digital ad options.
Now, Google ads get nearly 4-5% click-through rate (CTR).
On Instagram, ‘Feed Ads’ have a CLR of between 0.22% and 0.88%. ‘Stories Ads’ have a CTR between 0.33% and 0.54%.
Are you kidding me? And they are making $1bn per week?
We need new models for digital ads. But we also need smarter ways to think about digital privacy around ads, and how these platforms make money.
Lisa LeVasseur is the founder of Internet Safety Labs (ISL). She posts regularly about the devils in details of online privacy, including around digital advertising.
The ISL has reviewed hundreds of privacy policies, many of which agree that under CCPA/CPRA, behavioural advertising is "selling data”. Yet for Instagram, it’s covered under
"Personalizing the Meta Products"
"Sharing of Information across Meta Companies"
"Providing Ads on Meta Products"
Lisa rightly argues ‘How we make money’ should be an addition to the privacy policy.
Too right. Who’s doing that already?
The Unwinnable Cost Problem of Identity Verification
This is David Kelts’ second blog on the issues of ‘IDV’. This time he dives into the twin issues of cost and process around digital ID. And how we need to separate out two very important, but different, steps:
Document authentication - is it from a trusted issuer?
Document validation - is the data on it accurate and is the account behind it still valid?
He goes on to say that both get solved with digital government-signed identity data. So we can stop ‘ripping’ physical documents, which are both costly and imperfect.
Scanning plastic and paper documents has been a good stop-gap. But new smart digital approaches are now available. They are objectively better, cheaper, faster, more reliable and more private.
David offers four approaches that solve the cost and process problems of digital ID today. All involve portable, verifiable credentials.
He’s right that ID is an unwinnable cost problem.
So shout it loud folks: we urgently need new, public digital trust infrastructure that includes individuals as active participants, not just passive subjects.
Also known as ‘EmpowermentTech’.
Patient data is traded without us knowing - it can’t be solved, can it?
“An inability to control data-sharing can lead patients to withhold important information from their own health care providers.”
Wow things are messed up in the US around patient data.
In their latest report on Your Body But Not Your Data, research teams at RAND have been exploring the issues around personal data sharing in US healthcare. And the impacts and risks.
For example, it’s a massive effort for every researcher to ask every patient about every study. So consent for anonymised data use is normally granted retrospectively. The issue is that deindentified data can be traded - sometimes sold - without a patient consenting or even knowing.
Now over a third of patients are uncomfortable with their health data being used for research. Even if it's ‘anonymised’.
Just imagine the medical advances - and patient impacts - that become possible when patients can trust the digital tools they are using. And trust the organisations they are sharing data with.
Right now the data is not flowing because of digital trust.
This is fixable people. We have the data. We have the tech. We have the governance.
We just need to crack the incentives and use cases. Who’s working on this?
Who is going to dominate in identity wallets? It’s the economics, stupid
Last week I joined the excellent SSI Orbit podcast talking about Empowerment Tech.
We covered everything from personal agents and data regulations, to how the market is taking shape and some of the more interesting use cases.
But I want to point to you a part of the discussion near the end, around 47 minutes in. It’s a really important conversation about where data gets stored. About who might win. And why.
The punchline: it’s not about the protocol. It’s about the economics at the edge.
I’ve been following Sangeet Paul Choudry for years. He wrote ‘Platform Revoluion’ back in 2016, and is widely considered to be one of the leading writers on the economy and tech. He was recently selected as one of the top 10 publications ever published in the Harvard Business Review, alongside Michael Porter and Clayton Christensen.
If you are building a digital wallet, personal vaults or personal AI I strongly recommend reading a few of his latest posts. Including
Gen AI companions and the fight for the primary interface
When augmentation, not automation, is the real threat
The ONDC conundrum: Where protocols win… and where they don’t…
They cover the market dynamics for data. On how AI is changing the competitive landscape for businesses. And why - and how, and where - data is going to flow differently.
The third one in the list above is the most important. And is why I point to it in the NorthernBlock podcast.
Understanding the economics at the edge - of digital wallet and verifiable credentials protocols - will help you see who will win.
Spoiler: it’s about who can dominate at data storage, device-based AI and the economics of digital trust.
Biometrics are not credentials - and here’s what we can do about it
Mark Bennett reminds us of the important difference between biometrics (a core part of who you are) and credentials (which are assigned to you).
Credentials can be reissued, stolen, replicated, and shared. But biometrics are virtually impossible to steal or replicate.
Unless, as Mark points out, you give them away. What… like in a photo or video posted online? Folks, the toothpaste is already out of the tube.
There are three big challenges around biometrics:
Where should we store them? How should we think about security, privacy and data governance?
What happens when AI can replicate/fake them at near-zero cost?
What happens when bad actors get hold of them?
It’s where Empowerment Tech becomes a triple-fix.
Move biometrics to the edge
We first assume that ‘data at the centre’ is unsafe and unprivate (new word, I kind of like it).
Instead, we use our trusted devices ‘at the edge’ to manage our biometrics. Keep our sensitive data in trusted execution environments (TEEs) inside the hardware we already carry around.
We can then use trusted authentication steps to match the person using the device to what’s held inside (eye, face, finger, palm). The device becomes a trusted root for our digital wallets and personal agents.
And if you want to steal my data, you have to come after my face and my device. Not that centralised honeypot of my face and voiceprint, which can later be relayed to the bank to drain my savings account.
Move to intrinsic signals and credentials
We assume our ‘extrinsic’ biometrics are now foiled. Our gait, our smile, our voice can now be replayed and injected into someone else’s identity processes (like at the bank above).
Instead, we use a mix of ‘intrinsic’ signals and proof of biometrics from our devices. Patterns, behaviours and reputation data points from across our lives that are harder to reproduce. Together with harder-to-access biometric signals like palm and vein readers, which need a live pulse.
The win? We won’t need biometric readers everywhere. We just need the ability for digital wallets to share those private credentials, not the biometric itself.
Move to proof of ‘Customer-Present’
Today, credit card networks use ‘CardNotPresent’ signals to mean it’s a riskier payment transaction. I am not in the store physically, so it could be someone else using my credit card. And so the card networks charge more to cover the fraud.
We can do the same with our customer credentials. ‘Customer-Present’ (CP) Credentials can now provably come from our digital wallet, authenticated by our on-device biometric (as above).
Customer-NOT-Present (CNP) Credentials might represent my ‘real data’ (like my mother’s maiden name and DateOfBirth), but carry a much higher risk because it could be a 3rd party injecting them into an authentication process (like at the bank).
With this approach, businesses can begin to separate out those fraud signals. And lower the costs for the more trusted ‘Customer Present’ interactions.
In the end, bad actors can have my data. Much of it is already publicly available anyway. But it won’t be trusted unless they can prove that it’s a Customer Present transaction.
It’s all part of the EmpowermentTech story that’s going to drive fraud down, improve customer experiences and will be more compliant.
Mark Bennet is right. We need to separate out how we think about biometrics and credentials.
But whatever the answer, biometrics should live where they belong. At the edge with the individual, under their control.
How does Humane’s AI Pin work?
Humane has released a short explainer on how their ‘AI Pin’ works. It also digs into Cosmos, their AI Operating System.
It’s helpful to see how they think about weaving together different digital services across voice, vision and ‘knowledge’. Together they create Pin’s ‘memory’.
It describes how information moves into and out of Pin. And where it will take multiple ‘hops’ between providers (and how this introduces delays). I do wonder if some of this is counter-marketing, to fend off those throwing mud on performance…
The interesting thing though is where they want to take the AI Pin.
“The AI Pin can have a throughline through all of your computers. That means, for the benefit of these intelligent systems, your AI Pin will one day be able to interact and drive all of your computers.”
He’s describing a Personal Agent. ‘All of your computers’ won’t just mean a laptop and tablet. It will of course include
Your Digital Wallet
Your Personal Vault
Humane are pointing squarely to EmpowermentTech.
Right now, the AI Pin is mostly a smart interface for email, phone calls, reminders and Siri-style internet searches. It’s intelligent and cleverly designed.
But it feels like the iPhone1, which didn’t have copy-paste. And remember, iPhones didn’t really take off until iPhone3, two years later.
Directionally, we can see three things coming together
AI capabilities in the cloud (performance of GPT5, Claude and others)
Autonomous Agents (Large Action Models, or LAMs)
On-device compute (so we can move much more to the edge)
AI Pins, or devices like them, may soon organise much of your day.
The question, of course, is will they be trusted?
OTHER THINGS
There are far too many interesting and important Customer Futures things to include this week.
So here are some more links to chew on:
Talk: Decent primer on the EU Digital Identity Wallet from OpenID Japan WATCH
Article: Andreessen Horowitz on Putting Data in a Corner: LLMs and Financial Services Data READ
Blog: GTM in Identity and Fraud: Why Point Solutions Win READ
News: For the SSI nerds amongst you, TBD has announced a new DID protocol: DHT READ
Post: Regulation matters more than ever (because good product teams want to make products that are delightful, accessible and that reduce harm) READ
Article: New Laws Mandate Access To Social Media Data, But Obstacles Remain READ
And that’s a wrap. Stay tuned for more Customer Futures soon, both here and over at LinkedIn.
And if you’re not yet signed up, why not subscribe:
Hi Jamie, loving the newsletter and will try to attend your next meet up at Somerset Hse. I'd be interested in your thoughts on the intersection between smart data initiatives and their fundamental requirement to be based on 'Digital Verification Services' TF proposed in DPDI. Under the hood its all data so why replicate trust frameworks for the 7 sector-based schemes set out in the recent SD roadmap. Smart Data cannot happen without digital identity. Best wishes Nick, Assuriant Consulting.