I Failed on Amazon for Four Years. Here's What It Taught Me About SaaS Marketing.

It’s a Tuesday night in 2015, somewhere past midnight, and I’m staring at a spreadsheet I have looked at a thousand times before. Same columns. Same formulas. Same product. But the number at the bottom, the one that actually matters, has been moving in the wrong direction for months, and tonight it finally says something I can no longer argue with.

My margins are thinner than they were when I started.

Three years of work. Three years of optimization. I have rewritten every listing, tested every keyword, lowered my advertising cost of sale, negotiated my supplier down twice, chased reviews, watched my conversion rate climb. By every operational metric I have gotten better at this. And I am making less money per unit than I did on day one, when I knew almost nothing.

I sat there doing the thing every founder does at that hour. One more tweak. Maybe a new bid strategy. Maybe a new variation. Maybe a price test. I had a dozen levers and I had pulled all of them, and the floor kept rising toward me anyway.

That night was the beginning of the end of a business called Olely. And strange as it sounds, almost everything I do now as a marketer, I learned from watching it die.

What Olely Was

Let me pull back.

In 2012 I launched a private label cosmetics brand on Amazon. The category was beauty and personal care, the model was direct to consumer through the world’s largest marketplace, and the thesis was the one everybody had at the time: Amazon was a gold rush, and the people who moved fast would own shelf space before the crowd arrived.

I did everything right. I am being precise, not modest, because the failure matters more if you understand it was not a failure of execution. I sourced good product. I built a real brand with real packaging and a point of view. I learned Amazon SEO before most sellers knew it existed, ran PPC campaigns that were tight and profitable, built a review strategy that kept us credible, and optimized the listing down to the last bullet point. The whole playbook, run as well as almost anyone in my category.

And it worked, for a while. Then it slowly, quietly, stopped working.

The competition got fiercer. New sellers undercut me on price every quarter. Amazon changed its rules, its review policies, its fee structure, the layout of the page itself, and every change moved a little more leverage to Amazon and a little less to me. Advertising costs climbed steadily, because the only way to stay visible was to pay for it, and everyone else had figured that out too. I was running harder every month to stay in the same place. By 2015 I was running harder just to fall behind slightly slower.

It took me four years to admit the obvious. I have spent the years since turning that obvious thing into the way I think about marketing. Here is what Olely taught me.

Lesson 1: A Channel You Don’t Own Will Eventually Own You

This is the big one, and it is the one I missed completely while I was living inside it.

On Amazon, I did not own anything that mattered. I did not own the customer relationship, Amazon did. I did not own the customer data, the email, the purchase history, the ability to reach a past buyer, Amazon did. I did not own the traffic, the page, the rules of the page, or the terms on which I was allowed to keep selling. I was renting all of it. And the landlord could raise the rent whenever it wanted, change the locks whenever it wanted, and let a hundred competitors set up stalls next to mine whenever it wanted. It did all three.

When the algorithm shifted or the ad prices spiked, I had no leverage. None. I could not pick up the phone. I could not take my customers somewhere else, because they were never my customers. I had built a business on land I did not own.

That is the quiet injustice of rented channels, and it is not unique to Amazon. Builders pour real effort into a platform, and the platform keeps the relationship, the data, and the leverage. You do the building. The channel banks the equity.

I watch SaaS companies do the exact same thing now, and the structure is identical. When I work as a fractional CMO for SaaS businesses, the first risk I look for is channel concentration. A company that gets every lead from Google Ads is one CPC increase away from an unprofitable business, and they do not control the CPC. A company whose entire pipeline comes from organic SEO is one algorithm update away from silence, and they do not control the algorithm. They feel like they have a marketing engine. What they actually have is a tenancy.

The fix is not to abandon rented channels, because the lesson is easy to overlearn. Paid search, marketplaces, organic, these are powerful and you should use them. The fix is to build owned assets in parallel, deliberately, from the start: an email list you control, a community that knows your name, a brand reputation that travels by word of mouth, direct relationships with the people who buy from you. That balance, rented accelerants against owned foundations, is what real channel strategy is actually about. The rented channel is fine as an accelerant. It is dangerous as a foundation. Olely was built entirely on a foundation I was renting, and when the rent went up there was nothing underneath.

Lesson 2: Margin Is a Marketing Decision

Here is the second thing I got wrong, and it took me embarrassingly long to see it as a marketing problem at all.

Olely competed largely on price, in a category that was savagely price competitive. I had a brand, but I had not built a brand premium. There was no reason a buyer comparing my product to three nearly identical listings should pay more for mine, so they didn’t, and I cut price to win the sale, and so did everyone else, and that is the race to the bottom. Whoever is willing to make the least money wins, right up until they make no money, which is roughly where I ended up.

For years I thought margin was a sourcing problem, or a logistics problem, or a fee problem. It was a marketing problem. Margin comes from positioning: from giving a specific buyer a reason that your product is the only acceptable answer to their problem, so that price is not the deciding factor. I never gave anyone that reason. I gave them a good product at a fair price in a sea of good products at fair prices.

SaaS does this constantly. Companies lead with “cheaper than the incumbent,” and it works, briefly, until someone shows up cheaper than them, because there is always someone willing to burn more money for share. Competing on price is a position anyone can take from you. The companies that win compete on value, on being the obvious choice for a particular buyer with a particular problem, and that is a marketing decision made on purpose, not an accident of the spreadsheet. Positioning is the whole game. It decides your margin before a single ad runs.

Lesson 3: Fail Fast, Learn Faster

Four years was too long.

I should have made the hard call after two. By the end of year two I had the data. The trend was visible. The margins were compressing, the channel was tightening, the structural problems were already clear to anyone willing to read the spreadsheet without flinching. I read it and flinched, every quarter, for two more years.

The reason is the most common founder bias there is, and I had it badly: one more tweak, one more optimization, one more test, and surely this turns around. Optimization is seductive precisely because you are good at it and it produces small wins, and small wins feel like progress even when the entire model underneath you is broken. But sometimes the answer is not optimization. Sometimes the model is the problem, not the execution, and no amount of brilliant execution fixes a model that does not work. The bravest and most useful thing you can do is name that early and act on it.

This is, honestly, half of what I do now as a fractional marketing leader. When I look at what a fractional CMO does week to week, a lot of it is the willingness to say “this isn’t working, here is what we do instead” after sixty days, not after two years. An operator inside the business has every incentive to keep tweaking. The value of bringing in someone from outside is that they have no ego invested in the current plan and can call it early. The cost of waiting too long is not abstract to me. It is four years of my life and a business I loved, spent proving a point I already knew by year two.

The Real Lesson: Scar Tissue Is an Asset

Here is the thing nobody tells you about failure. The best marketing lessons I have do not come from the things that worked. They come from Olely, from watching something I built well die anyway, for reasons that had nothing to do with how well I built it.

But scar tissue is only an asset if you are honest enough to study it. Plenty of founders walk away from a failure with the wrong lesson, or no lesson, because looking at it clearly is painful. The whole value of the scar is in the diagnosis: not “Amazon is hard,” which is useless, but “I built on a channel I didn’t own, I never created a margin through positioning, and I waited two years too long to act.” Those three sentences are worth more than any success I have had, because a success rarely tells you why it worked. A failure, studied honestly, tells you exactly why.

That is what I bring to every engagement now. Not a playbook of things that worked once, but the willingness to look at a business and call the hard thing early: the channel risk nobody wants to name, the positioning that does not exist yet, the thing that is not working that everyone is too invested to stop doing. I learned all of it the expensive way, so that the founders I work with do not have to.

Back to the Spreadsheet

So what did I finally do, that night in 2015 and the months after it?

I stopped optimizing and started telling the truth. I admitted the model was broken, not the execution. I wound the business down instead of feeding it for another two years, and I took the lessons and went and got good at the thing those lessons were actually about, which was marketing strategy: the part that sits above the tactics, where you decide which channel, which position, and how long before you call it.

If I could go back and sit next to the version of me at that desk past midnight, I would not give him a new bid strategy. I would close the spreadsheet. I would tell him the number is not lying and no tweak is coming to save it, that he is renting a business he thinks he owns, that he never built the margin that would have made any of this survivable, and that the hard call he is dreading is the cheapest thing on the table tonight and gets more expensive every month he waits.

And that is exactly what I would tell a founder staring at their own version of that spreadsheet right now. The channel that feels like an engine might be a tenancy. The growth that feels like traction might be a race to the bottom. And the optimization you are about to spend another quarter on might be the most expensive way there is to avoid the decision you already know you need to make. Read the number. Then decide what kind of business you actually want to be building, and whether you own it.