Why Growth Is a Product Problem
The most durable growth comes from embedding distribution into the product itself — not from marketing campaigns that fade.
The clearest lesson of my career came from a comparison I did not plan to run. At AliExpress US we were doing two things at once: spending heavily on paid acquisition, and building an affiliate growth engine that eventually reached more than 10,000 partners across 40-plus countries. The paid spend produced revenue in a straight, predictable, entirely rented line. The affiliate engine produced roughly a 30 percent revenue lift and, critically, kept producing after the team stopped adding partners.
The difference was not channel economics. It was that one of those things was a campaign and the other was a system embedded in how the product worked. Campaigns decay the moment you stop feeding them. Systems compound.
That is the whole argument, and most growth organizations are structured to ignore it.
Marketing rents demand. Product manufactures it.
I want to be precise about what I am claiming, because product-led growth has been diluted into meaning almost nothing.
I am not saying marketing is useless. Paid acquisition is a perfectly good instrument for buying a specific outcome at a known price, and if you have a working funnel and available capital, you should use it. What I am saying is that paid acquisition produces no asset. When the spend stops, the growth stops, and you own nothing you did not own before. Every quarter you start at zero and buy your way back to where you were.
Product-led distribution works differently because the mechanism lives inside something you own. When we built the creator commerce program at AliExpress, onboarding around 400 creators a month and reaching more than 2,000 active creators for about 40 percent GMV growth, the growth was not coming from the campaigns we ran to recruit creators. It was coming from the fact that the product made creators successful, and successful creators recruited other creators and kept producing content without further prompting. The recruiting campaign was the ignition. The product was the engine.
Here is the test I apply. Ask what happens to this growth if the team responsible for it goes on vacation for a quarter. If the answer is that it stops, it is a campaign. If the answer is that it slows but continues, it is a product mechanism. Almost everything labeled growth in most organizations fails this test.
The four places distribution actually lives in a product
When people hear product-led growth they usually think of a free tier or a viral invite loop, which is a SaaS-shaped view that does not transfer well to commerce. In my experience distribution gets embedded in one of four places, and only one of them is the famous one.
- The output is shareable. Using the product produces an artifact that travels on its own. A creator's video, a public wishlist, a comparison someone sends a friend. The growth is a byproduct of normal usage.
- The product makes a partner more successful. Affiliates, creators, resellers, and integrators promote you because doing so pays them better than the alternative. This is the mechanism that scales widest in commerce and gets the least product investment.
- Value increases with participation. Reviews, fit data, sizing feedback, compatibility reports. Each user makes the product better for the next, which is why review depth is a moat and not a feature.
- The product is discoverable by construction. Structured data, catalog completeness, and content that answers the question a buyer actually typed. This is unglamorous and permanently undervalued.
The second one is where I have spent most of my career, and where the gap between rhetoric and investment is widest. Almost every consumer brand says partners are strategic. Almost none of them staff a product team against the partner experience. The affiliate program gets a dashboard built in 2016, a monthly CSV, and payouts that arrive whenever finance gets to it.
Then the brand wonders why its partners prioritize a competitor. The answer is usually that the competitor built a better product for partners, and it was not close.
Why organizations resist this
If the argument is this straightforward, the obvious question is why more teams do not act on it. I think there are three honest reasons, and none of them are stupidity.
The first is attribution. Campaign growth is measurable this week. Product-led growth shows up in a curve that bends over two or three quarters, and by then six other things have changed. If your growth team is evaluated monthly, it will rationally choose the thing it can prove.
The second is organizational placement. Growth usually reports into marketing, which means the growth team can commission a campaign but has to negotiate for engineering time. Given a choice between a lever you control and a lever you have to beg for, you pull the one you control. This is not a talent problem, it is an org chart problem, and it is fixable only from above.
The third is that campaign work feels like progress. There is real psychological comfort in shipping a campaign every two weeks. Building a partner platform means three months where the number does not move and someone senior asks what you have been doing. Most people are not willing to absorb that, and most managers are not willing to defend it.
If your growth team reports into marketing and has to requisition engineering time, you have already decided that growth will be campaign-shaped. The org chart makes the strategy, not the strategy deck.
What it looked like in practice
At Indiegogo I worked on the playbook for taking hardware campaigns from crowdfunding into real ecommerce, across more than 100 campaigns. The pattern was consistent and instructive. A campaign would raise well, and the founders would conclude they had product-market fit and a growth problem. They almost always had it backwards.
What they actually had was a one-time distribution event. The crowdfunding platform had supplied the audience, and that audience did not belong to them. Founders who understood this spent the next six months building mechanisms that generated their own demand: content that ranked, a referral structure with genuine economics, a reseller motion, community that produced usable content. Founders who did not understand it spent the same six months buying ads to replace an audience they thought they had earned.
The second group's numbers looked better for about a quarter, which made the first group's decision harder to defend internally. That lag is the tax on doing this correctly, and there is no way to avoid paying it.
The same pattern held at Next2Market across consumer electronics brands including Samsung, TCL, Levoit and Anker, where we drove 350 percent GMV growth and an 18 percent conversion rate improvement. The conversion improvement did not come from better ad creative. It came from fixing product surfaces that were quietly destroying intent: incomplete specification data, comparison experiences that forced people to leave, checkout flows with unnecessary steps. That work is indistinguishable from product work, and it was accounted for as growth. The labels do not matter. The ownership does.
How to tell if you are actually doing this
Four diagnostics, in order of how uncomfortable they are.
- What fraction of your engineering capacity went to growth mechanisms last quarter? If it rounds to zero, your growth is rented regardless of what your strategy says.
- Can you name a distribution mechanism that got better without anyone working on it this month? Compounding is the whole point, and if nothing compounds you have a very expensive treadmill.
- If you cut paid spend by half for one quarter, what is your honest revenue forecast? Most teams have never modeled this and are frightened to.
- Do your partners, affiliates, and creators have a product experience anyone on your team would describe as good? Go use it yourself. The answer is usually no, and usually nobody has checked in years.
The reason I care about this beyond efficiency is that product-led growth produces better alignment. When distribution depends on the product genuinely working for someone, you cannot grow by being better at persuasion than you are at delivery. A campaign can outrun a mediocre product for a long time. A referral loop cannot. A creator program cannot, because creators stop promoting things that disappoint their audience.
That constraint is a feature. It forces the organization to earn its growth rather than purchase it, and the growth you earn is the only kind you keep.