How Loomhaus Grew GMV From $310K To $1.74M In 8 Months

When Loomhaus first came to me, the numbers told a story that anybody running a two-sided marketplace will recognize instantly, and it is a brutal one. They were sitting at roughly $310K in monthly gross merchandise value, they had about 1,200 active sellers who had listed at least one product in the trailing ninety days, and they had a buyer base that was technically large on paper at around 48,000 registered accounts, but the repeat purchase rate was a miserable 9 percent, so the entire business was basically a leaky bucket that the founder was refilling with paid acquisition every single month. The blended take rate was 14 percent, which means that $310K in GMV was only throwing off about $43,400 in actual marketplace revenue, and out of that they were spending almost $31,000 a month on Meta and Google acquisition to keep both sides of the network from collapsing, so the contribution margin after acquisition was painfully thin, and the founder told me flatly that at the current burn they had maybe seven months of runway left.

The deeper problem, and this is the thing that almost nobody diagnoses correctly, was that they had a content vacuum sitting right in the middle of a two-sided cold-start problem. Sellers would not commit listings and inventory to a marketplace that could not prove it drove demand, and buyers would not develop a habit around a marketplace that felt thin and impersonal and that they only ever saw through a retargeting ad, so both sides were waiting on the other side, and the founder was paying cash to manufacture the appearance of liquidity that the content was failing to make real. Their entire content operation at that point was one overworked person posting maybe three product flat-lays a week to Instagram, the seller-side blog had not been updated in five months, the YouTube channel had four videos with a combined 900 views, and there was no founder presence anywhere, no operator voice, no point of view, nothing that a maker browsing for a marketplace to join could latch onto and trust.

The acquisition math was the part that kept me up at night when I was scoping this. Their blended customer acquisition cost across both buyers and sellers was sitting at $61, and a new buyer was only worth about $34 in contribution margin in the first ninety days, so they were underwater on every single acquisition and they were betting the whole company on a lifetime value that the 9 percent repeat rate said was never going to materialize. They had tried an agency before me that promised performance creative, they had burned $24,000 over a quarter on a library of generic ad variations, and the result was a 0.7 percent click-through rate and a CAC that actually went up, so by the time we talked the founder was deeply skeptical of anyone promising content would fix a numbers problem.

So the challenge I signed up for was not really make better videos, the challenge was prove that organic, distributed, platform-native content could lower acquisition cost on both sides of a marketplace at the same time, lift repeat purchase behavior on the buyer side, pull in higher-quality sellers on the supply side, and do it fast enough that the runway math flipped from terrifying to boring, and the founder gave me exactly eight months to show it in the GMV line and not just in vanity reach.

My whole model is the flywheel, and I explained it to the founder on the first call exactly the way I am going to lay it out here, because it only works if the client understands that the leverage is in distribution and not in volume of production. The flywheel is one shoot a month, and that one shoot gets cut into thirty-plus platform-native assets, and those assets get distributed everywhere both sides of the marketplace already spend their attention, and because they are native to each platform they compound instead of disappearing, and the compounding is what makes qualified leads, both buyers and sellers, arrive already warm instead of stone cold off a retargeting pixel. That is the entire thesis, and for a two-sided marketplace it is doubly powerful because one shoot can feed both the demand side and the supply side if you architect the narrative right.

So the first thing I did, before a single camera came out, was sit down and rebuild the content architecture around the two audiences, because a maker deciding whether to list on Loomhaus needs a completely different proof than a buyer deciding whether to check out, right, and the mistake the previous agency made was treating it like one funnel. I built two narrative spines. The seller spine was operator-to-operator, the founder talking candidly about how the marketplace actually drives demand, real seller payout numbers, real conversion data, the kind of transparent inside-the-business storytelling that makes a skeptical maker think this is a place run by people who get it. The buyer spine was discovery and taste, the maker behind the product, the story of the object, the why-this-costs-what-it-costs honesty that turns a one-time buyer into someone who comes back because they trust the curation.

Then we set up the monthly shoot as a single production day that I scoped to feed both spines at once. Every month we would shoot the founder doing two or three operator-voice pieces to camera for the seller side, and we would shoot a featured maker or two for the buyer side, and we would capture the b-roll, the product detail, the process footage, all of it in one tightly run day, and that single day is what generated the thirty-plus assets. From one shoot I would cut a long-form YouTube anchor piece, three or four mid-form pieces for LinkedIn and YouTube, eight to twelve short vertical cuts for Reels, TikTok, and Shorts, a written operator essay for the seller blog and the email list, a buyer-facing maker-story post, ten-plus static and carousel pieces for Instagram and Pinterest which matter enormously for a home-goods marketplace, and a batch of native ad creative pulled straight from the organic winners so the paid amplification was riding content that had already proven it resonated.

That last part is the connective tissue I want to be clear about, because the financial story depends on it. I do not treat organic and paid as separate departments, I let the organic distribution find the winners for free, and then I take only the assets that actually earned attention organically and put a modest amplification budget behind them, so the paid spend is never guessing, it is always pouring fuel on something the audience already validated, and that is the mechanism that took their blended ROAS from a money-losing position up to 6.2 to 1 over the engagement. We were not spending more, in fact by month four we were spending less on paid than they had been before, we were just spending it on creative that the flywheel had pre-qualified.

Distribution was deliberately wide and deliberately native. On the seller side that meant LinkedIn for the operator content where founders and makers actually take business seriously, YouTube for the long-form trust-building anchor, and a re-activated email sequence to the dormant seller pipeline. On the buyer side that meant Instagram and Pinterest carrying the visual weight because that is where home-goods discovery happens, TikTok and Reels for the maker-story shorts that travel, and a buyer email program rebuilt around the maker narratives instead of discount blasts. Everything pointed back to specific marketplace category pages and specific seller storefronts, so the reach was never just brand awareness floating in the void, it was reach engineered to deposit a warm visitor onto a page that could convert them.

And I instrumented the whole thing from day one, because the founder had been burned and I needed the financial line to be undeniable, so we tracked GMV, take-rate revenue, blended CAC split by side, repeat purchase rate, seller activation rate, pipeline value of inbound seller applications, sales-call volume from the higher-tier seller outreach, and the close rate on those calls, and we reviewed it every two weeks against the eight-month target so there was nowhere for a vanity metric to hide.

Let me give you the financial picture the way I gave it to the founder in the final review, because this is the part that actually matters and everything else was just the work that got us here. We started at $310K in monthly GMV and we closed month 8 at $1.74M in monthly GMV, and that is a 5.6x increase in the single number that defines a marketplace, and it did not happen in a spike, it happened as a clean compounding curve, $310K then $412K then $538K then $694K then $902K then $1.18M then $1.46M then $1.74M, and you can see in that sequence that the growth rate held because the content library kept compounding rather than resetting every month.

Now translate that into the money that hits the business. At a 14 percent blended take rate the starting GMV was throwing off about $43,400 a month in marketplace revenue, and by month 8 the same take rate on $1.74M was throwing off roughly $243,600 a month in marketplace revenue, so we added almost exactly $200,000 in monthly recurring marketplace revenue over the eight months, and because the take rate actually crept up to 14.8 percent on the premium seller tiers we converted, the real number was a touch higher than that. The cumulative GMV across the engagement summed to about $7.24M, which against a baseline trajectory of flat $310K months would have been roughly $2.48M, so the content flywheel is directly responsible for something on the order of $4.76M in incremental GMV inside the eight months alone, and the evergreen library is still producing after the engagement closed.

The acquisition cost story is the one I am proudest of, because this was the founder's existential problem and we solved it without spending more, we solved it by spending smarter. Blended CAC across both buyers and sellers started at $61 and ended at $29, a 52 percent reduction, and that happened because warm traffic from compounding organic content converts at a fundamentally higher rate than cold retargeting, so every paid dollar stretched further and a large slice of new acquisition simply arrived organically at zero marginal cost. Blended ROAS on the paid amplification went from a money-losing 1.4 to 1 up to a healthy 6.2 to 1, and remember we were spending less on paid by month four than they had been spending before the engagement, so the ROAS gain was pure efficiency from riding organically validated creative.

The repeat purchase rate, which is the metric that decides whether a marketplace is a real business or a leaky bucket, went from 9 percent to 24 percent, and that is a 2.7x improvement in the single behavior that compounds lifetime value, and it happened because the maker-story content gave buyers a reason to come back that had nothing to do with a discount, it built actual taste-based loyalty to the curation. On the supply side the premium seller close rate went from 18 percent to 41 percent, more than doubling, because by the time a maker got on a sales call they had already consumed weeks of operator-voice content that pre-sold them on the marketplace, so the call was confirming a decision instead of making one.

The funnel for the eight months tells the same story in volume. We put 4.38 million in monthly reach at the peak, we converted that into about 657,000 engaged interactions across the period, those produced 41,200 qualified leads split across buyers and inbound seller applications, that generated 412 sales calls with the premium seller pipeline, and we closed 169 of them into premium listing relationships, and each of those premium sellers carries a materially higher take rate and a materially higher GMV contribution than the baseline seller, so the close at the bottom of the funnel was disproportionately valuable.

On the bottom line, the total engagement investment was $152,000 across the eight months covering production, distribution, and the modest amplification budget, the pipeline value we generated in projected annual take-rate revenue from the warm seller applications and buyer cohorts crossed $3.9M, the directly attributable closed revenue inside the window was $1.76M in marketplace take-rate revenue, and that puts the return on the engagement at 11.6x, and that number is conservative because it does not even credit the evergreen library that keeps compounding after we stopped, right. The runway problem that the founder walked in with, the seven months of cash left, is simply gone, because the contribution margin per acquisition flipped from negative to strongly positive somewhere around month four and never looked back.

I want to pull back from the numbers for a second and talk about why this worked, because if you are a marketplace founder reading this, the temptation is going to be to copy the tactics and miss the mechanism, and the mechanism is the whole thing. The reason one shoot a month could move GMV by 5.6x is not that we are magicians with a camera, it is that a two-sided marketplace has a content-shaped hole right in the middle of its cold-start problem, and almost nobody fills it correctly because they treat content as marketing decoration instead of treating it as the liquidity-manufacturing engine that it actually is.

Here is the thing about a marketplace that the spreadsheet never tells you, the sellers are waiting for proof of demand and the buyers are waiting for proof of selection, and both of them are sitting there with their arms crossed waiting for the other side to move first, and the founder is standing in the middle burning cash to fake the liquidity that would make either side commit. Paid acquisition can buy you a buyer for a day, but it cannot make a maker believe in your demand, and it cannot make a buyer develop a habit, so the cash you spend on ads is buying motion without belief. Content, distributed natively and allowed to compound, is the only thing I have ever found that manufactures belief on both sides at the same time, and belief is what flips the flywheel from something you push to something that spins on its own.

That is why I architected two narrative spines from the same shoot day instead of one. The operator-voice seller content was doing a very specific job, it was letting a skeptical maker watch the founder talk honestly about real payout numbers and real conversion data, and that transparency is worth more than any pitch, because makers have been burned by marketplaces that overpromise, so honesty is the unlock. And the buyer-facing maker-story content was doing a completely different job, it was turning anonymous transactions into relationships with the people behind the products, and that is what took the repeat rate from 9 to 24, because you do not come back to a faceless catalog but you do come back to a curator whose taste you have learned to trust.

The distribution discipline is the part people underestimate the most. It is not enough to make thirty assets, you have to put each one where it is native, because a LinkedIn operator essay dies on TikTok and a maker-story Reel dies in a seller email, and the reason the reach compounded from 142,000 to 4.38 million is that every single asset was built for the exact surface it lived on, so the platforms rewarded it instead of suppressing it. And then the financial leverage came from refusing to separate organic and paid, because I let the free distribution find the winners and only then put money behind the proven ones, so the paid budget was never a gamble, it was always an amplifier on validated demand, and that is the entire reason the ROAS went from 1.4 to 6.2 without spending a dollar more.

The last thing I will say is about durability, because a lot of growth is rented and disappears the day you stop paying, and this growth is owned. The evergreen library we built over eight months is still on YouTube and Pinterest and the blog doing its job, still pulling in warm sellers and habit-forming buyers at zero marginal cost, so the $1.74M monthly GMV is not a peak we have to defend with ever-increasing spend, it is a floor that the compounding content keeps lifting. That is the difference between buying attention and building an asset, and it is the whole reason a boutique distribution agency running one shoot a month can outperform a performance shop burning ten times the ad budget.