You Can't Tailor Network Effects

Most ecommerce “platforms” aren’t platforms at all. They’re aggregators in ready-to-wear suits, mistaking scale for connection. Everyone calls their software a “platform” now. It’s become one of those words that means everything and therefore nothing, like “synergy” or “quiet luxury” (which I loathe, in case you missed that).

Shopify is a platform. BigCommerce is a platform. Your email service provider is probably calling itself a platform. That Chrome extension you installed last week? Also a platform, apparently.

But here’s the thing: most of them aren’t. They’re aggregators cosplaying as platforms, and the difference matters more than you think.

If you understand why, you’ll understand something fundamental about how value actually gets created in networks. And also why most ecommerce businesses are optimizing for the wrong things.

Metcalfe’s Law (The Thing Everyone Cites and Nobody Understands)

Robert Metcalfe, who invented Ethernet and clearly had time to think about networks, came up with a deceptively simple formula: the value of a network is proportional to the square of the number of users.

Or more precisely: V = n(n-1)/2, if we’re being pedantic about it (which I am, because I use fountain pens and am a stickler for good manners, so pedantry is sort of my brand).

Side note: I did read a rather interesting refutation on Matcalfe’s Law that you may find refutable in itself.

But the idea is simple. A telephone network with one phone is useless. Two phones? One connection, marginally useful. Ten phones? Forty-five possible connections. A hundred phones? 4,950 connections. The value explodes as you add users because each new user can connect to everyone already there.

This works for actual networks. Telephones. Fax machines (remember those? I am deeply and forever in love with the principle and practice of the humble fax machine). Facebook, unfortunately. The more people who have them, the more valuable they become to everyone.

Classic network effects. Beautiful, elegant, and almost never what’s actually happening in ecommerce.

Second side note: we’re not chatting about blockchain and Bitcoin though Metcalfe’s Law is often applied to Bitcoin and blockchain networks because their utility and value increase as more participants join and interact. However, it’s not a perfect fit because network value in blockchain also depends on transaction volume, utility, and trust, not just the number of users.

Reed’s Law (For When Metcalfe Isn’t Ambitious Enough)

David Reed went further. David Reed argued that Metcalfe’s Law (n²) only accounts for pairwise connections, while in group-forming networks such as online communities, social platforms, or blockchain ecosystems where users can create countless subsets or groups. Each new user can potentially join or form multiple combinations of these groups, so the potential value grows exponentially as 2ⁿ.

Which grows a lot faster. Exponentially faster, technically speaking.

This is why Slack works. Why Discord works. Why any platform that lets people form groups and communities becomes genuinely valuable. Each possible group configuration adds value.

But Shopify? BigCommerce? Your ecommerce “platform”?

Not so much.

The Ecommerce Platform Delusion

Here’s an uncomfortable question: does your Shopify store become more valuable because there are more Shopify stores?

No. Obviously not.

You don’t benefit from other merchants existing on Shopify. You’re not more successful because someone else opened a store. You’re not connected to them in any meaningful way. Their success doesn’t help you. Their failure doesn’t hurt you.

You might both use Shopify, but you’re not in a network with each other. You’re in an aggregation.

Shopify is selling you infrastructure. Very good infrastructure, to be fair. But infrastructure isn’t a platform in the Metcalfe sense. It’s an aggregator that creates value through economies of scale, not through network effects.

The distinction matters.

Real vs. Fake Network Effects

Let me give you some actual examples of network effects, since we’re apparently all confused about this.

eBay has network effects. More buyers attract more sellers. More sellers attract more buyers. Each side makes the other side more valuable. That’s a two-sided marketplace with actual network effects.

Visa has network effects. Merchants accept it because customers have it. Customers get it because merchants accept it. Classic network effect.

Facebook has network effects (unfortunately). You’re on it because your friends are on it. Your friends are on it because you are. Each new user makes it marginally more valuable to everyone else.

Shopify does not have network effects. Your store isn’t more valuable because there are more stores. The app marketplace has weak network effects at best (more merchants might attract more developers, but it’s not exponential growth, it’s linear at best).

Shopify creates value through aggregation and economies of scale. They can negotiate better payment processing rates because they aggregate demand. They can invest in better infrastructure because they have more customers sharing costs.

That’s valuable. That’s a good business model. But it’s not a platform with network effects.

It’s an aggregator pretending to be a platform.

Why This Matters (Beyond Semantic Pedantry)

I realize this sounds like I’m being unnecessarily precise about definitions (I am), but the distinction has real implications.

Pricing implications: If something has genuine network effects, you should be willing to pay more as it scales because you’re getting more value. If it’s just aggregation, you’re paying for economies of scale, which should make it cheaper over time, not more expensive.

Shopify’s pricing doesn’t reflect network effects. It reflects operational costs and value captured from economies of scale. That’s fine. But it means you’re not getting compounding value from other users existing.

Competitive dynamics: True platforms with network effects are winner-take-all. Facebook beats MySpace beats Friendster because network effects create a moat. Once everyone’s on Facebook, switching costs are massive.

Ecommerce platforms? Not winner-take-all. You can switch from Shopify to BigCommerce to WooCommerce without losing any network value because there was no network value to lose. Your customers don’t care what platform you’re on.

EXCEPT THEY DO - BUT THAT’S ANOTHER ARTICLE. STAY TUNED.

Defensibility: If you’re building something that depends on “platform effects,” you better make sure you actually have them. Most ecommerce businesses don’t. They have aggregation benefits, which are valuable but different.

The Coordination vs. Aggregation Distinction

Here’s the useful framework, which I’m stealing from Ben Thompson but making my own through the power of restating it with more fountain pen references:

Platforms coordinate. They create value by enabling connections between users. More users = more potential connections = exponentially more value.

Aggregators consolidate. They create value by reducing costs through scale. More users = better economies of scale = linearly more value.

Most ecommerce “platforms” are aggregators. They consolidate demand to negotiate better terms with payment processors, hosting providers, app developers. That’s useful. But it’s not the same as coordinating connections between users.

The confusion happens because both can be called “platforms” in casual usage, but they work completely differently.

If you’re building on an aggregator while thinking it’s a platform, you’re making strategic mistakes. You’re expecting exponential value growth from linear scaling.

(I’m writing this with my Sheaffer PFM V with the 18k Stub Nib, which I bought because of network effects - sort of. I read reviews, watched videos, asked people in fountain pen communities. The community creates information sharing that approximates network effects, even though fountain pens themselves don’t have them. (See? Everything connects if you think about it sideways.)

Side note three: not actually writing this with my beloved Sheaffer, but I did scribble down the opening line in a notebook that had been personally bound for me by my dear friend, which counts.

What Actual Network Effects in Ecommerce Would Look Like

So what would genuine network effects look like in ecommerce?

Marketplaces have them. Etsy, eBay, Amazon (the marketplace, not the retailer). More buyers attract more sellers. More sellers attract more buyers. That’s a real two-sided network effect.

Payment networks have them. Stripe, PayPal, Klarna. Merchants want them because customers have them. Customers have them because merchants accept them.

Loyalty programs might have them. If your loyalty points work across multiple merchants, there’s a network effect. More merchants accepting the points makes the program more valuable to customers, which makes more merchants want to join.

But your Shopify store? Your email list? Your abandoned cart recovery tool?

No network effects. Just good infrastructure and maybe some economies of scale.

Why I Think About This While Building Something

I’m building something in the ecommerce software space (deliberately vague, you’ll find out when I’m ready to tell you). And one of the core questions I keep asking is: where are the actual coordination problems vs. where are the aggregation opportunities?

Most ecommerce “platforms” are trying to aggregate. Consolidate merchants, negotiate better rates, create economies of scale. That’s fine, but it’s not where the interesting problems are.

The interesting problems are coordination problems. Getting marketing, dev, and merchandising to work together. Connecting work to outcomes. Making sure everyone knows what’s happening without requiring seventeen status meetings.

Those are real network effects, in a way. The more your teams are coordinated, the more valuable coordination becomes. Each additional connection between team members, between tools, between work and outcome, creates exponential value.

That’s a network effect worth building for.

The Uncomfortable Implication

If most ecommerce platforms aren’t actually platforms with network effects, what does that mean for you?

It means you’re not locked in the way you think you are. Switching costs are lower than platform rhetoric suggests. You’re not getting value from other users, so leaving doesn’t cost you network access.

It means you should probably pay less than you’re currently paying, since you’re not getting exponential value from scale.

It means the competitive dynamics are different than venture capitalists want you to believe. There’s no inevitable winner-take-all outcome. Multiple “platforms” can coexist because they’re not actually platforms.

And it means if you’re building something, you should be very honest about whether you have network effects or just economies of scale. One is exponentially more valuable than the other.

(I’m somewhat precocious, in case that wasn’t clear by now.)

What This Has to Do With Bespoke Tailoring

You knew I’d get here eventually.

Bespoke tailoring has no network effects. My tailor doesn’t become better at his job because more people commission suits from him. In fact, he probably gets worse because he has less time per client.

But there’s a community effect that approximates network effects. People who appreciate craft talk to each other. They share recommendations. They form groups on Reddit and discuss lapel widths and trouser breaks.

That community creates information flow that helps everyone make better decisions. It’s not Metcalfe’s Law, but it’s something.

Ecommerce platforms keep trying to create those communities (merchant forums, app ecosystems, “partner networks”), hoping to manufacture network effects where none naturally exist.

It rarely works. Because you can’t fake emergent coordination through aggregated infrastructure.

You either have network effects or you don’t. And most ecommerce platforms don’t.

They’re just really good at aggregation and really good at calling it something else.

###

I write about economics, ecommerce, and why words matter more than people think. Sometimes about the best way to bake a cheesecake and the fact that it’s actually simple to tie a bowtie. Currently building something in the software space.

PS: If you disagree about any of this, especially the network effects analysis, I’d genuinely love to hear why. I collect data on being wrong. It’s surprisingly useful.