Zero Clicks #17: Damned lies and Black Friday statistics

Every week in Zero Clicks, we explore the interplay of AI, media, and commerce.

This week we’ll dive into…

• Big picture: Dueling Black Friday data from Adobe & Salesforce (and why Shopify’s report is the only one worth paying attention to)

• Job posts: LEGO just dropped a late contender for ‘Dream Job of the Year’

• The meme is the message: The end of affiliate arbitrage?!

The Super Bowl of commerce is in the books and according to Adobe Analytics, American shoppers spent $10.8B online this Black Friday. The problem is, nobody really knows how they came up with that number, and seemingly nobody has really bothered to ask.

According to Adobe, the data is “based on 1 trillion visits tracked to U.S. retail sites, covering some 100 million SKUs and 18 product categories in all.”

Cool, sounds fancy. But here’s what it actually means. Adobe Analytics does some kind of guesstimate extrapolation of overall sales volume based on the fraction of true Black Friday transaction volume that they see from merchants who run Experience Cloud. They put out a press release with this number and it gets printed reliably in NPR, Modern Retail, and ABC News, with few questioning if it is a reliable representative sample.

Trust issues

This dance has been going on in our business for more than a decade. As an added bonus, reporters also always mention that Black Friday sales broke records because yeah, that’s kinda how growing economies with inflation work.

It’s nearly impossible to fathom that Adobe can publicly share any Amazon sales data and Adobe only sits on a fraction of Shopify sites and some FORTUNE 1000 retailers. So what are they looking at exactly? Just an extrapolation from enterprise retailers and brands on Magento and BigCommerce? That's uh....not exactly where the action is these days in eCommerce.

Meanwhile, Shopify, which does actually have a complete first-party data set on GMV across its platform, reports that consumers spent $5B on Black Friday this year, a 22% YoY growth rate, nearly double what was reported across the mainstream press. While this is partially a story about Shopify taking market share, it’s also the only real complete data we have for proper year-over-year comparison.

Shopify’s $5B number for volume on its platform is global while Adobe’s $10.8B estimate of volume across commerce is only the US market.

For its part, Salesforce estimates that US customers spent $17.5B online during Black Friday, a casual $6.7B difference from Adobe, though they curiously only estimate 7% YoY growth from 2023. Salesforce’s claim is based on “shopping data from 1.5 billion consumers captured across its customers and other data feeds in its Commerce Cloud, Marketing Cloud, and Service Cloud.”

I don’t really know what the hell that means but I do know that Salesforce also only sees a fraction of online commerce (that doesn’t include Amazon) and has its own indecipherable method of extrapolating from a small base. That I happen to think they are closer to the truth than Adobe misses the point. Neither of these numbers should be cited by the mainstream press without Barry Bonds’ level Hall of Fame asterisks.

There are three kinds of lies: lies, damn lies, and statistics

The broader problem is that the annual dump of questionable Black Friday data is anything but an aberration. All sorts of borderline made up numbers dominate media headlines across all types of issues. These generally come in two forms.

First, there are the comically egregious claims from seemingly very important authority figures– McKinsey’s claim that the metaverse would generate $5T by 2030 is the piece d’resistance of this genre. But Gartner, Accenture, and BCG get in on the nonsense, too. Then there’s the obvious bullshit, like the notion that Gen Z employees feel they need a $580K salary to be successful. Thankfully while grifters eat this shit up, the mainstream press generally is savvy enough to avoid publishing this nonsense.

Far more common are the kinda, sorta accurate numbers like Adobe’s that are directionally correct but presented with dishonest faux precision and often cited by respected outlets. The most famous example of this is the wildly ubiquitous stat that 60% of Americans live paycheck to paycheck, which is based on a single study by Lending Club which refuses to release its methodology. Do you think a lending company might be just slightly incentivized to imply that a lot of Americans need emergency loans?

For years in commerce, one of the most commonly referenced statistics around Amazon’s dominance is that 55% of consumers begin their product searches on Amazon. This number was first published in 2018 by Bloomberg, CNBC,  Vox, Business Insider and spread to dozens of other outlets thereafter.

But here’s the rub– this seminal eCommerce stat was based on a single, flawed survey from a tech vendor named Bloomreach that was sent to a grand total of 2,000 people.

Here’s how the question that led to this statistic was originally posed to respondents:

When shopping online, where do you generally begin your search for products?

• Amazon

• Search engine (i.e. Google) 

• A retailer’s website  

Now, let’s imagine the same question posed with a more extensive set of choices.

• Amazon

• Google 

• Perplexity

• Instagram

• Pinterest

• Temu

• Tiktok

• Directly on a retailer’s website

• I exclusively shop in physical stores 

Just by adding a few more options, you’d get a wildly different dataset, to say nothing of other behavioral economics factors like whether or not the order of different choices is randomized. Bear this in mind when you see confident prognostications that Google is for old people or TikTok is the pre-eminent search engine of our era.

Even if survey methodologies are solid, important context on where data comes from inevitably gets lost in the game of hyperlink telephone that plays out on the internet. Here’s how it generally goes. The first news outlet that publishes a third party statistic will often make a cursory attempt to cite the methodology and link to the original survey. But then the second outlet simply publishes the stat and links back to the first news story. Pretty soon, a given statistic is accepted as gospel without any clear acknowledgment of how it was derived.

With malice towards none but laziness by all, dodgy statistics enter the mainstream.

Check your sources folks and unless a number is clearly based on first party data, take it with a gargantuan grain of salt.

GM, Enterprise, Substack

While power laws are just starting to really eat commerce, they’ve already devoured the creator economy, which is a business of haves and have nots with a miniscule middle class.

Substack seems to have realized this reality in a hurry, understanding that it needs to win further upmarket if it is going to grow into its $600M+ valuation. The company is now looking for a leader to do just that, presumably in the form of building tools for large enterprise publishers and premiere creator media operations to run their businesses on Substack.

On the surface, entering the crowded CMS space (or honestly, selling publisher tech broadly) is not the most attractive market but Substack might have the moxie to do something special here. For a talented media product mind, this is an epic gig.

Sr, Director, Head of Adult Engagement, LEGO

I have no pithy commentary here other than to say that this job is absolutely adorable and I envy the hell out of whoever gets it.

Account Executive, Advertising Sales, StockX

On a long enough timeline, everything inevitably becomes an ad network. At least Stock X’s will have a hell of a lot of swag. Great opportunity for a junior media seller to accelerate their career in retail media. 

In honor of whatever the hell Google is currently doing with its site reputation abuse policy, I present this meme.

To think that it all started because this guy got a bad CBD gummies recommendation from Forbes…

Disappointed that only one person correctly identified that last edition’s mystery stock at all-time highs was the New York Times Company.

So naturally, let’s make it harder. Here’s the 5-year price chart for a retailer whose stock rose through the bear market of 2022, touched all-time highs in 2023 and 2024, and has been down this year as the market goes up. Name the mystery company.

Reply directly to this email with your guess. First one to guess correctly (without help from AI) gets a shoutout next week (plus unlimited bragging rights in their company Slack).