Ads Broke News and Why Substack Won't Save It

The subscription model is the latest panacea for all media's woes, with Substack having just raised its $65M series B from Andreessen Horowitz . With journalists complaining for the past ten years that their revenue from ads isn't enough to sustain essential bodily functions, why deal with the big bad admen when you can?ask your readers for money??Diogenes is smiling from his wine jar in heaven.

Controversial take: enabling better subscriptions is only a minuscule part of fixing our media economy.?The question of, 'hmmm, advertising models seem to be working pretty great for Google and Facebook (henceforth?called?GooBook), why isn't it working for publishers' is still largely misunderstood due to the opaqueness of how online advertising works.?

And we can only fix news if most understand what's breaking it.

"Subscriptions are the Future! Advertising is Dead!"

In my journey through the media meta-verse, many tell me ads are dead, and subscriptions are the future.?Hearing this causes me physical pain.

Firstly, subscription models and advertising models are not mutually exclusive. They've coincided for centuries:

1. Humans carved something on stone, and this stone was then sold to someone who wanted it.
2. Then, we invented the printing press, which enabled mass production and the scale needed for the subscription model.
3. Then, some guy (Benjamin Day ) realized he could print cash by plopping ads on his newspapers.

And in 2021, our media meta-verse ('subscriptions are the future') suggests that an evolutionary regression from 3 → 2 is the future.?Congratulations, Darwin is crying.

Though it may seem slightly reductive, I'm not saying that pure-play subscription models can't ever work — it works well for many. I'm stating that subscriptions and ads have played nice for centuries and that relying solely on one mode of revenue generation over another doesn't capture the entire consumer surplus of a publisher's product (news).

In plain English, relying solely on subscriptions is a backward step in realizing print news's total economic and social value.

A High-Level Primer on How Modern Digital Ads Work

So we all want a sustainable fix for our media economy, and logically, pure-play subscriptions aren't it. News has two revenue streams: subscriptions and ads — which seem pretty broken. So, to fix ads, we must first understand how they work.

Good thing I'm here to tell you!

Most digital ads are traded on real-time exchanges (like stocks!). Every time you click on a web page, a piece of inventory is added to that site's ad 'warehouse,' which is then auctioned off to thousands of advertisers, with the winning bid being the ad you see. All this happens in milliseconds, and like stocks, ad prices are constantly in flux and directly relate to how valuable advertisers think an ad is in driving sales for them.

More effective ads = more bids = higher prices!

Like investment bankers, advertisers like to make ads sound complicated with their CTRs, CPMs, and XYZs. But if you look through the jargon and understand the first principles, it's like any other marketplace.

Why Ads Are Broken (for publishers):

Platform Centralisation Powers Ad Efficacy and Boosts Market Prices (for centralized platforms)

Long story short, ads are broken for publishers as publishers need platform centralization. Platform centralization in digital media is the ownership of the content (e.g., the News Feed) and ad delivery mechanism (Facebook's Ad Marketplace). Combining these mechanisms enables ads on Google Search and Facebook to fetch prices up to 5x that of advertisements on digital publications .

These efficiencies include:

• Better targeting (by collecting more + higher-quality data about you) and,
• Better UX (when was the last time you remembered a banner ad? That's due to something called banner blindness , something solved with more excellent user journeys like Instagram stories),

Platform centralization by enabling better targeting and better UX makes ads more useful for advertisers, leading to higher demand, thus pushing their prices up on the 'ad stock exchange'.

This is why advertisers are willing to pay this 5x premium and why GooBook earns the big bucks: one can reasonably conclude that advertisers see GooBook ads as 5x more likely to convert into a sale, giving advertisers an excellent Return on Ad Spend.

Advertisers are willing to bid up and pay 5x more for one ad on Google Search and Facebook as they're better at driving sales due to the power of platform centralisation enabling better targeting and better form factors.

Remember when I said that every time you load a page, a piece of inventory is added to a warehouse and then auctioned off to advertisers? As publishers are small and building warehouses is complex, publishers rent their warehouses from another company instead. This means publishers own the content delivery mechanism (WordPress) and monetize their content via Google Display Network (which powers 80%-90% of the display). This architecture is fundamentally decentralized, so publishers are technologically barred from having the above advantages.

N. b., because of how exchange-sold ads perform, most publishers circumvent the exchange and sell ads directly to advertisers, fetching higher prices. However, even more is needed — I've included an analysis of the annual reports of Facebook and the Daily Mail below.

Assuming the Daily Mail sells 100% of its inventory directly, Facebook would still earn 2. 5x more per user hour. This analysis also shows the value of a brand. Assuming the 5x GooBook ad price multiplier applied to the Daily Mail, we have a theoretical '100% exchange sold ad' revenue generation of ~70￠/hour. This means at the price of $1. 45/hour, advertisers value the 'halo effect' of the Daily Mail brand at a ~2x premium over the base cost of the ad itself. More on this later.

Note: Unlike most ad-revenue analyses, this analysis uses hours spent on the platform rather than page views, and this approach is more accurate and complicated. For the curious, you can find the method and calculations here:

Google is Incentivised Against Bringing These Efficiencies to Publishers

So, if Google takes a cut from nearly all ads on publications, it makes sense that Google is incentivized to help publishers lift their ad revenue. Unfortunately, the opposite is true.

Ads on publications and search exist in indirect competition — the lower price of publication ads enables a far grander scale, and the brand value of publishers allows ads on magazines to be especially suited to creating brand awareness. GooBook's ads, being far more targeted (but commanding far higher prices), make them more suitable for direct conversions.

Hence, while these two ad types may not be in direct competition, if publisher ads were one day to gain the targeting and UX of GooBook, they would be competing in a segment with far?higher margins.?Bearing in mind that advertisers can value a publisher's brand an extra 2x the base cost of the ad itself, one can tentatively conclude that doing this could take significant market share in the market segment currently dominated by GooBook.

And herein lies the problem: for every publication ad sold, Google has to share 80% of revenue with publishers, and for every Search Ad, Google keeps 100%.?This means it's?imperative to Google's core business to prevent publisher ads from laddering up into the market segment currently occupied by its Search product.

Imagine you're Google. I don't want to share 80% of my revenue with publishers! Think of the shareholders!

"The smaller the banner ad pie, the bigger my search pie, and the less 80% revenue share I need to give to publishers! And I literally control both markets! Genius!" - Google (probably)

Here's a graph showing US Newspaper Ad Revenue (source: Newspaper Association of America). In 2007, Google acquired DoubleClick, the world's largest ad server for digital publications. While this comparison may seem fallacious due to the chart illustrating print and digital news revenue, if GooBook were included, they would exceed this graph's bounds by over 500% (basically a vertical line).

Hence, the takeaway is this: as consumers transitioned from physical news to digital, the market size for advertising has quintupled in a span of 15 years, which Google, for all its talk of supporting publishers, has essentially helped publishers take 0% of.

So holding up subscription models as the solution for publisher woes is like applying a Band-Aid on a limb turned into meat gravy by a Google-branded IED — gravy Google is turning into fresh pot pies and pairing with freshly decanted publisher tears (yum).

Like burning coal, industry concentration in ad tech causes significant economic externalities by suppressing publisher ad revenue, meaning publishers have less to spend on quality journalism, which leads to shittier news and a weak fourth estate.

When viewed this way, the reality is that subscription models are a compromise publishers have made due to their inability to raise ad prices.

To Conclude

To solve the problems facing news, we need to address the underlying problem of how tech giants use their market power to limit ad prices for publishers.

There are two solutions. Either:

• The break up of Big Ad-Tech (GooBook) happens, making digital ads terrible for everyone (again) or
• News media unify under a shared publishing platform, which enables both the centralized platform advantage and (unlike brandless platforms like Medium) allows them to build a brand; in doing so, step up to compete with big tech (like what Shopify did with eBay).

The first option is a terrible idea, which might seem surprising as I just compared Google and IEDs.

Regardless, the second option is the more interesting, being as of yet unexplored. Unifying print media under a shared, centralized platform disrupts current ad tech players and ancient publishing groups like NewsCorp, which have stayed relatively unchanged since their start. Instead of unifying publishers under a publishing group, why don't we decentralize publishers under a shared software layer that powers better ads?

Note: In a prior version of this article, I calculated brand value uplift via a comparison between pageviews of MailOnline and Facebook. While this approach is simple, we can get an even more precise economic comparison if we compare both platforms 'aggregate user time spent on the platform to determine the cost of attention. Using this analysis, we can see that the brand value of DailyMail provides a 2x, not 40%, uplift over the base cost of the ad itself.