∞

Stanley is trying to dial back the drops frenzy by borrowing a page from the Taylor Swift playbook and getting people to sign up for the chance purchase a new product when it drops.

What goes into the screening process?

browser signals, location signals, fraud analysis, cohort signals and “assessing basically who looks like a real fan and who is using automation or otherwise to game the system,”
…
A red flag, for example, would be when the system detects that far too many people are entering from one IP address.

A lot of data science to sell an insulated straw cup.

via CNN

∞

Another Podcast’s episode Google Gemini and AI bias unpacks the weights and training data stuff I briefly mentioned yesterday.

Hidden patterns in data and models as bias amplifiers.

Also the excellent quote “the fake is in the caption,” because what is a prompt but a caption for a future image?

∞

I’m a big pedal steel fan, so when Jason Isbell started talking about it on the Broken Record podcast I took note.

The pedals are so small, and they’re so close together, that if you don’t wear the right kind of shoes you’ll accidentally hit more than one pedal at a time. So this is why pedal steel is a true cowboy instrument, because cowboy boots are what you wear in order to be able to hit just one pedal at a time.

Pedal steel (a key country music instrument) also has a learning curve.

What are the pedal steels and cowboy boots of what you’re trying to do? Are you using the right ones?

∞

From episode three of A History of Rock Music in Five Hundred Songs:

Once you give a collection of things a name—the way people’s minds work—they start thinking because those things share a name, they’re the same kind of thing.

Names and genres and labels matter.

The human brain is designed to pattern match and categorize and bucket and metaphor to save on processing power.

(Neural networks are designed to do this for AI models via weights.)

In grocery aisles, “water company” usually means plastic bottles—it’s what shoppers expect. Liquid Death smashes that expectation, so it stands out.

∞

Finished reading: The Last Horse: Prologue by Aaron Tucker 📚

∞

Finished reading: John Constantine, Hellblazer Vol. 8: Rake at the Gates of Hell by Garth Ennis 📚

∞

The more you know 💫

Ada Lovelace, the daughter of poet Lord Byron, wrote the first algorithm created specifically for a machine.

-Tom Asacker

∞

In the section of a recent The Email Marketing Show episode titled “Don’t make the mistake of offering discounts upfront!”:

You got to really legitimize discounts and bonuses and things like that and you do that by really planting that seed and creating that familiarity.

Don’t introduce people to your brand via sales or promos (unless it’s a major shopping holiday where a sale is expected).

∞

Investment is (slowly) trickling back into ad tech

recent activity is a stark contrast to the conservatism of investors in 2023, with speculation centering upon the AI and CTV sectors.

AI is not a surprise, slap that label on anything and you’ll attract investment right now.

CTV is the heir apparent to the traditional TV advertising throne, even if YouTube and other creator-content based platforms will carve a sizable slice of that pie for themselves.

∞

Ambient computing is coming

Apple has explored the idea of developing new wearable devices — including a fitness ring, smart glasses and even AirPods with cameras — to broaden one of its most important business areas.

This is a mix of Apple targeting products from other brands (Oura, Meta Ray-Bans, Snap Spectacles, etc) and building the product ladder that leads to the Vision Pro as the eventual replacement for the Mac line.

via Bloomberg

∞

What happens when the usual forms of human-computer interface get disrupted?

OpenAI is working on AI agents that take over your mouse and keyboard, performing the tasks in real time.

  • One would complete complex tasks on your device like creating a spreadsheet from a document of information or filling out your tax forms.
  • The second agent would take on web-based tasks like curating data from different sites (similar to Perplexity or Arc Browser), booking flights and hotels, or even building travel itineraries.

The groundwork has been laid, AI could be the accelerant.

via The Future Party

∞

Wendy’s looking to test surge pricing at restaurants

The Amazonification of physical retail.

I’m interested to see how this goes. I don’t think it’ll be received well by consumers. I don’t think people want to be surprised by their burger price while ordering.

∞

This is maybe not the thing to say in a time of inflation when It’s Been 30 Years Since Food Ate Up This Much of Your Income:

We’ve got to reach the consumer where they are, so we’re advertising about cereal for dinner. If you think about the cost of cereal for a family versus what they might otherwise do, that’s going to be much more affordable.

-CEO of Kellogg | (non-video link)

It does not matter if this is factually accurate or not, your consumer’s aren’t pleased that this is where they’re at. And Frosted Flakes don’t exactly align with the surging wellness trend.

∞

after the spike in online shopping ignited by the 2020 Covid quarantine, the e-commerce share of all retail sales had shrunk to less than 15% in the third quarter of 2023. Over the past four years, the e-commerce slice of retail industry revenue has gained by only 0.5%.

But, as this Forbes piece, ecommerce is becoming more and more enmeshed in the full retail ecosystem. It is no longer siloed off.

Apps power in-store shopping. In-store returns trigger on-site purchases. Ecommerce discovery turns into in-store purchases.

Just like most things digital these days, it’s omnichannel.

∞

According to the Dotdash Meredith CEO:

“if you cannot show performance, you are dead.” The evolution of advertising and how people interact with those ads has caused a shift to bottom-up priorities, instead of top-down.

via Sounds Profitable

∞

Spotify has big goals for its ad product

The long-term aspiration, as stated by CEO Daniel Ek, is for advertising to generate 20% of the company’s revenue. The longer-term goal on top of that, according to Couchman, is reaching $10 billion in ad sales.

via Sounds Profitable

∞

Some quotes from Dave Gerhardt on the Marketing Against The Grain podcast:

You say small business, mid market, and enterprise, internal. External that often means nothing.

The way you speak about what you do matters. Use your customer’s language, not your internal business speak.

People buy products when they say, “oh, interesting, this company looks like mine, I’m going to use that product.”

This is making the case for case studies.

Or, as Seth Godin says: People like us do things like this.

∞

From a great Dan Oshinsky LinkedIn post on email metrics:

What does a good click-to-open metric look like? In 2023, ConvertKit said the average CTOR was 9.2%, while Mailerlite said their average was 8.9%. If you’re beating those numbers, that’s a very positive sign!

Click-to-open is an underrated metric.

∞

In The Abstract: Effectives of Limited Time Promos in Online Retail

The paper Examining the Efficacy of Time Scarcity Marketing Promotions in Online Retail from the Journal of Marketing Research has some great research (as covered on the Nudge Podcast).

Even the abstract is informative, if a bit unwieldy in that classically academic way:

Time scarcity promotions (e.g., “40% off for a limited time”) are mainstays of online retail marketing. Although positive effects of time scarcity promotions on consumer interest have been evidenced in the brick-and-mortar world, should retailers expect similarly robust effects online? The present research suggests the answer may be no. First, the authors report meta-analytic and experimental results suggesting that previously identified positive effects of time scarcity promotions observed offline may not emerge in online shopping contexts. Then, consistent with the prediction that online time scarcity promotions activate more persuasion knowledge than identical control promotions, the authors detail findings suggesting that providing retailer-exogeneous justifications for online time scarcity promotions’ time restriction (e.g., consumers’ birthdays, seasonal changes) can increase the potential of observing positive effects on consumer interest online. Further, results suggest that the positive effects of including exogenous time justification may be more likely when less time remains until the online promotion’s expiration. However, results stop short of suggesting that online time scarcity promotions will consistently yield superior outcomes compared with identical online control promotions. Therefore, the authors highlight the continued need for careful managerial use as well as further research examining the optimal translation of offline tactics to online retail.

I will now try to translate this into actual human speak, line by line…

Limited time promos / flash sales are popular with ecommerce.

These types of offers have proven success in brick-and-mortar / offline retail, but does that transfer online?

Spoiler alert: probably not.

This is based on a review of pervious research.

Providing a reason for the promo’s limited run can boost the offer’s performance online (e.g., holidays, celebrations, seasons, etc).

This can be boosted further by shortening the offer’s time frame (e.g. 24 hours only).

But these steps won’t make them super promos when compared to other offer types you might run.

We recommend you use these carefully. And more research is needed on the performance of offline tactics for ecomm.

A few guesses as to why flash sales and limited time promos aren’t as effective online:

  • Offline limited time promos have the added constraint of physical presence / travel. There may be 3 days left on a sale, but can you get back to this store in the next 3 days? Our brains might even process the sale as being positive ROI on our time investment that got us into that store (unsubstantiated hypothesis).
  • Snaps and Stories may have altered the timeline of the internet to 24 hours. Anything longer than one solar day is no longer urgent.
  • The internet is an infinite shelf of content, decisions are made in terms of “should I do this now?” 72 hours left in a sale means it doesn’t need to be shopped RIGHT NOW, and is then likely to be forgotten before the expiration date.
  • Grounding the promo in something outside the feed or brand’s business goals can stand out as a hook. “25% for 24 hours” doesn’t have a hook (discounts and promos are ubiquitous now, they aren’t hooks in and of themselves). “25% off on your birthday” might trigger “I should treat myself.' “Spring Sale - 25% off” might trigger “I could use some pieces to freshen things up.”
∞

The New York Times is reportedly developing a generative AI tool to enhance ad targeting. The national daily newspaper is experimenting with several large language models (LLMs) to power the new tool, which will be capable of matching ads with ideal consumers based on their interests, goals, and opinions.

AI will turbocharge contextual advertising.

via Inside

note: I did not find reference to advertising in The Verge article Inside cited. So I don’t know where the above claim is really from, but it does make sense.