Now that I’m well into the fourth year of running this publication, it’s time to make a decision. I spend a lot of time creating the newsletter and it’s no longer sustainable to do so for free, in all honesty.
I also have a lot more I would like to offer readers, if I could dedicate the time to hi, tech.
For instance:
Primers and guides on topics like generative AI, data privacy, TikTok Ads, and much more
Interviews with interesting people in tech
Audio and video content
Trends webinars and events
Community discussions (the Substack app just added great features for this)
With that in mind:
🔥🤖 Hi, tech. will offer a paid subscription, starting… now.
I will still send out a free post every week, offering a snappy look at the big stories in tech, marketing, and AI. However, the detailed analysis will all be paid from now on. Historical posts will also be behind the paywall for members only.
You can sign up monthly (£10/month) OR I’m offering a 20% discount on the regular annual fee of £100 right here:
There are also group discounts available if you sign up through a work email address, so it’s worth rounding up any insight-hungry colleagues.
I love sharing this content with you all every week and I really hope to continue doing so. It’s time to build on what hi, tech. has achieved so far and make it an even more valuable resource for the tech and marketing community.
I appreciate your support very much - see you on the other side!
hi, tech. is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.
There's plenty of pain to go around. Companies across the industry are cutting costs, freezing new hires, and laying off staff. Employees who joined those hyped pre-IPO companies and took much of their compensation in the form of stock options are now deep underwater and can only hope for a future rebound.
This is fantastic. You’ll have seen the new Kia logo - it looks reasonably slick and modern, if you ask me. As it happens, it is too slick and modern for over 30,000 people every month. They’re searching for “KN car” instead of “KIA car”.
Hive is not a direct Twitter clone. Instead, the Gen Z-focused social app was described in a Teen Vogue profile last year as combining concepts from a variety of social networks, including both Instagram and Twitter and even MySpace — the latter thanks to a feature that lets Hive users add music to their profiles, among other things.
The app is also not solely timeline-based, as Twitter is. In addition to the main feed, Hive’s users can explore their interests across a range of topic-based communities.
And now for the first of our 2023 specials, and the last free one 😵
hi, tech. is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.
🔮 Digital Marketing in 2023
In which, I aim to make the following points:
We still don’t know the difference between the ‘selection effect’ and the ‘advertising effect’. That conflation was convenient in the past; it will be a hindrance in the future.
Yes, this all relates to Apple’s App Tracking Transparency (ATT), Google’s cookie changes, and data privacy laws.
Consider 2022 the prologue to the main story.
Digital marketing’s silver bullet has been data-driven targeting. We’re not great at the marketing basics (4P’s), nor are we very creative. Targeting has allowed us to excel while staying at the fringes of the enterprise. In the upcoming age, we’ll need to earn a place at the top table if we want to succeed.
We don’t even know how effective our targeting is today (see point 1), but we know it sort of works.
Targeting needs tracking. Take away the tracking and you take away the data. A new type of targeting means new types of data and a new type of strategy.
Everyone is jostling for position within the customer journey. Those who already own the point of purchase (Walmart, Tesco, etc.) have an inbuilt advantage and can build out from there.
Retail media is growing very quickly as a result.
They’ll get greedy and fall into the same traps as everyone else in this game: self-congratulatory reporting, murky data usage, too many ads in search results. Just because data is “first-party” does not mean it complies with data privacy regulations.
The above can be positive for the industry, although we have not laid the foundations we need to thrive. Digital marketing needs to get closer to the customer (not in that usual, creepy tracking way), calculate and test serious metrics, and be much more imaginative with content creation.
Have you ever seen a banner ad you liked?
Consider generative AI: We could train models to stimulate new ideas for content. We could use our customer insights data as an input to the models. But that entails an entirely different production process to syncing an ecommerce feed with a dynamic retargeting tool.
This all requires greater influence over the enterprise. That must be earned by doing more than offering vapid “search volume” trends as the “voice of the customer”.
Digital marketing needs to mature.
1: In 2019, a group of economists published an article that got people talking - briefly.
In reality, it should have had us all talking for a lot longer.
The authors had all worked on digital advertising systems for the likes of Google, Yahoo!, and eBay.
The crux of their argument was that:
“The benchmarks that advertising companies use – intended to measure the number of clicks, sales and downloads that occur after an ad is viewed – are fundamentally misleading.
None of these benchmarks distinguish between
the selection effect (clicks, purchases and downloads that are happening anyway)
and
the advertising effect (clicks, purchases and downloads that would not have happened without ads).”
Very few in the industry ever wanted to get to the heart of this matter, for fear of what they might find.
As an ex-eBay economist reflected on one meeting with ad executives:
"I looked around the room, and all I saw were people nodding their heads."
Let’s think of the implications of this wilful confusion.
We have:
The selection effect: A brand places an ad in front of a customer that was going to purchase anyway, then attributes the sale to the ad. The ad didn’t do much, but it takes the credit anyway.
And
The advertising effect: Advertising shapes a consumer behaviour and influences them to take an action they would not otherwise have taken. This, surely, is what advertising is about.
It is not even that difficult to work this out, if you run some experiments.
It’s what we teach people to do in our MBA course, primarily because we know so few businesses really take the time to get it right. The fact that the digital marketing modules are so popular in MBA courses should be cause for optimism, too.
But why am I focusing on a dry discussion from 2019 for my future-focused 2023 article?
Well, because the industry never tackled this question, we don’t know how effective digital marketing really is. We have self-congratulatory reporting from Google/Meta and a chorus of obedient mediocrities waiting impatiently to repeat the industry standard message they receive from above.
When everyone was making some money, the room was full of nodding heads.
This discussion will be brought to a head next year and brands will no longer be able to avoid it. The tide is going out, as the saying goes.
2: The defining characteristic of the digital marketing age is data-driven targeting.
Targeting is the practice of directing ads to groups of people who are predicted to interact with the advertiser’s products or services.
Targeting depends on tracking, which the World Wide Web Consortium (W3C) defines as,
“the retention, use, or sharing of data from [online] activity outside the context in which it occurred.”
Tracking is typically achieved through cookies, which enable servers to store information about users.
The co-creator of cookies, Lou Montulli, wrote in 2013 that “Tracking across websites was certainly not what cookies were designed to do,” but they proved an irresistibly effective method for personalising online advertising.
The majority (56%) of digital advertising spend in 2022 will go to display advertising, which uses technologies such as cookies to track consumers as they browse different websites. A 2019 study by Karaj et al. revealed that Google is the largest of these trackers by reach, with Google third-party scripts identified on 82% of monitored web pages.
No-one knows what will replace cookies yet, but we know for sure that regulators will no longer tolerate this kind of tracking.
3: As a result of point 2, publishers, social networks, and retailers are jostling for position in the customer purchase journey.
We therefore see the rise of “retail media”:
(This leaves out Apple, which will be a very significant player in digital advertising. We’ll tackle that in another special edition soon.)
Instead of tracking people across the web and then using this data to target them later, retailers want to maximise the value of their own corner of the web.
Walmart Connect offers everything from in-store activations to online banner ads and search marketing.
The idea here is that:
“Consumers are frequently logged in using their personal credentials when they shop online, allowing retailers to collate particulars about shopper interests without running into restrictions.”
Those who already own the point of purchase have the coveted data everyone else needs.
That does not automatically make them compliant with data privacy laws.
In fact, one potential eventuality for retail media in 2023 runs like this:
Retailers will do what businesses do: They’ll ruin a good thing by getting greedy.
They will run too many ads (Amazon is already on this track)
They will provide overconfident self-reporting (think ‘selection effect’ instead of ‘advertising effect’)
They will increasingly engage in murky data practices, such as using customer data for purposes other than those for which they have explicit consent.
I’ll repeat a hi, tech. truism: Marketers do not read data privacy regulations. They read blog posts that tell them the “3 Mega Takeaways You WON’T Believe” based loosely on the regulations.
Almost everyone overlooks this fact: There is no legal distinction between first-party data and third-party data. Just because you are on the Tesco or Walmart site and you are logged into your account, that does not mean they have free rein to do what they like with your data.
I am dealing only with retailers here, but in fact everyone will compete to own the customer journey. In the process, social media will become more like ecommerce, while ecommerce sites will add social media-esque “engagement” features.
OK well if you’re so clever, what’s the solution?
The above sounds more pessimistic than I actually am. However, as we know with Brexit here in the UK, you can only overcome problems if you address them directly. If you ignore these problems, they don’t seem to go away.
So let’s start with what we need and then we’ll work backwards to chart a path to the destination.
To thrive in this new age, effective digital marketing will:
Capture data thoughtfully: Instead of tracking users without their consent, we’ll have to explain up front what data we want, what we’ll do with it, and what the customer gets in return.
Influence the enterprise: If you are advertising on a retailer’s website, the ads will be more effective if you have some influence over product and price, as well as promotion.
Get closer to customers: That might sound strange, given the context. Yet I am not talking about following them here; marketers need to blend qualitative insights with their quantitative data.
Create more imaginative content: When we can no longer hound customers across websites, we’ll have to go back to the eternal truths of marketing. Cheap impressions cheapen content.
Analyse ROI with some sophistication: Rather than taking Google’s word for it that Google did a great job, we’ll test and figure out the real impact of our efforts.
To take advantage of this new landscape, digital marketing can no longer simply “optimise” at the fringes.
Today, the organisational structure might look something like this:
As a consequence, let’s see where digital marketing sits in the food chain:
The corporate strategy is set from above
Annual budgets are set between the CFO and CMO
The digital infrastructure is managed by I.T.
Important decisions about the product are taken elsewhere
Digital marketers take what they are given and deliver the best results they can. They have little say in product management, finance, I.T., or business strategy.
Again, we need to tackle the distinction between the ‘selection effect’ and the ‘advertising effect’ to gain credibility. We’ve spent years telling everyone how great our results are, without questioning the data too much. Why would they think anything needs to change?
I’ll tell you a story
I worked at a digital marketing agency and they won a contract with a well-known retailer in the US. Like many legacy retailers, this one was struggling a little in the digital age.
The agency provided consultancy and management for Google and Facebook services. A couple of months into the contract, there were media reports that this retailer had posted positive sales. The client services team at the agency sent an all-company email congratulating the account team for their brilliant work.
A year later, this retailer went into administration. Client services went quiet, thank heavens.
The point is not that the agency was either “good” or “bad” at their specialism. The point is that their efforts were largely irrelevant; they were rearranging the deck chairs as the Titanic headed for the iceberg.
They wanted to consider themselves “business partners” to this retailer, but what they offered was expert knowledge of how to work the advertising platforms. Even if they were exceptionally good at this, it would make a percentage point different here and there at best, versus a rival agency. It is commodified knowledge.
When I pointed out that the retailer’s issues were to do with their business model, their outdated products, their high prices, and their slow adoption of ecommerce, I was told that these were not the challenges the agency was contracted to solve. Admittedly, I wasn’t as gracious as I have just made myself out to be.
It was certainly true that we were doing what we were paid to do. And yet, we called ourselves a “digital performance agency”. How can you be a performance agency with access to none of the levers that dictate performance?
We repeatedly got lucky: Google or Facebook would launch new products, show us how to use them, then our results would improve. Or, Google and Facebook would tell us they had improved.
If we reintroduce one of the real classic marketing frameworks (4 P’s) we see just how superficial digital marketing can be:
Product
Price
Place
Promotion
Which of these do we cater to? Place, perhaps - we choose the channels most likely to connect with the right audience. We certainly cover promotion, too. The others have been lost along the way as we settle for a marginal role with an outsized (perhaps even overstated) impact on the bottom line.
If digital marketers assume the new role outlined above, most notably by establishing a thoughtful structure for customer data capture, they will be much better equipped to add value elsewhere. The rhetoric about being the “voice of the customer” could never be true while we were looking at flimsy data.
This means elevating the role of digital marketing and establishing formal connections with other departments:
That level of influence can only ever be earned by demonstrating value.
I am sure that if we put these foundations in place, digital marketing can be much more customer-friendly, creative, and collaborative. I’ve probably been saying that for over a decade now and each year proves me wrong, yet there are some reasons to be hopeful. Put bluntly, circumstances dictate that we have to make the transition this time.
This is before we even add in the phenomenal pace of progress in other areas of technology, such as generative AI.
Generative AI opens the potential for entirely new forms of creativity in marketing content. Instead of just sticking the images we get from the brand or e-commerce team into a social ad, we can train algorithms to generate unique images for our brand and our audience.
This is all very exciting but we are building on sand if we do not address some foundational issues first.
For example, if a brand has little idea of the true ROI of its marketing activity today, will generative AI help or hinder their measurement efforts?
Adding that level of intellectual rigour to our work will showcase the emerging maturity of the industry.
In the process, we can truly prove that the ‘advertising effect’ is the driving force behind our results.
