Darkroom

We look at over 500 consumer marketing P&Ls every year. The pattern right now is the same across almost all of them. Costs are up, efficiency is down, and nobody has a clear answer for what comes next. 1. The old DTC playbook is done. Facebook ads to landing page to conversion funnel worked for a decade. It doesn't now. CPMs keep going up, attribution is still broken, and most teams genuinely don't know where their next dollar should go. 2. Commoditization is now the default for most product categories. Since 2020, the number of Shopify merchants went from 1.75 million to 5.5 million. Competing on product alone is a losing game for most brands now. How you market, where you show up, and how fast you move is what actually separates people. 3. TikTok Shop, Amazon, and DTC are not separate strategies. They feed each other. What happens on TikTok affects what happens on Amazon. Brands treating these as siloed channels are leaving a lot on the table. You have to win in each one. None of them are optional. 4. Most teams still can't tell you where their profit is actually coming from. We look at over 500 consumer P&Ls a year. Margin is tight while budget is tighter. Almost none of the advertising programs we see are built to track actual profit. They track proxies. That gap is where most of the money gets lost. 5. There is no brand marketing vs performance marketing. That distinction is costing people. Everything is performance marketing. If it doesn't drive a measurable result, it doesn't matter. 6. Everyone will have AI. That's not the advantage. There's a small subset of people who actually understand how to use it in this industry. Not in theory. From experience, having built and run these programs. That's the scarce resource right now. It's a question of how fast you can get to them before everyone else figures that out. 7. Creative is still the main driver. That part never changed. Everything gets automated eventually. Media buying, targeting, optimization. Creative is still what makes people stop and care. CMO tenure has been declining for ten years straight. A lot of that is because the fundamentals got abandoned in the chase for efficiency. The craft still matters. The next two to four years will separate the companies that adapted from those that didn't. We know where the puck is going. We've built Darkroom to get there first.

Last year we hosted our first creator villa in the French Riviera. This year we’re back in France with the Darkroom CPG Garden in partnership with the legend Oren John at Cannes Lions International Festival of Creativity 2026. Our team is hosting an open day for the brands, founders, operators, and creators building what’s next in consumer and commerce. Blue sea, pines, palms, a long lunch and the conversations the Croisette doesn’t have time for. June 23, RSVP here: https://lnkd.in/dkpyc696

The mechanics of TikTok Shop are simpler than the hype suggests. Put your product in the hands of hundreds of people every month. Build in a reason for them to post about it. Let the algorithm, and human curiosity, do the rest. We hear it constantly from DTC brands and service providers: "Should we be on TikTok Shop?" But the question underneath that one is usually different. What they actually want to know is how to engineer the same kind of momentum for their own business. Look at what's happening inside the platform. The sampling cadence sits between 200 and 500 customers a month, each one nudged to share an honest reaction, what surprised them, what hooked them, who they'd hand it to next. You can't pay for that nor can you fake it. You have to build the conditions for it. The playbook for this isn't locked inside TikTok. Reverse-engineer the mechanic and it ports cleanly to Instagram, YouTube, and Meta, and usually performs harder when it runs across all of them at once. It comes down to two moves. 1. Get your product in front of people worth listening to. 2. Give them something worth talking about. 

Everybody goes into TikTok Shop hoping to go viral. I watched it happen from the inside as Head of US Haircare at TikTok Shop. Brands would blow up overnight. Creators organically loved their product. No ads needed. It was exciting. But that was then. The marketplace has become more saturated. The number of top creators has not grown at the same speed as sales. Those creators have become more coveted. And getting their attention to position yourself for a viral moment has become more challenging. What you actually need from a TikTok Shop partner is somebody who understands a formula that doesn't rely on virality. Someone who knows the inputs that drive the outputs and can speak to the speed of those inputs as they learn about your product. Because not all products are the same. Not all categories are the same. And even within the same subcategory your customer may adopt faster or slower depending on the messaging and how quickly it reaches enough people to drive conversion. Then there is the non-sexy stuff: How often are you refreshing your PDPs? what’s the CRM strategy? Have you tested multiple keyword strategies? Are your promotional strategies tied to traffic driving opportunities? These tactics drive higher profitability GMV to counter other platform costs. Virality is exciting. It's what everybody hopes for. But it isn’t a strategy, its vibes.

ROAS is a useful number. It's just not a strategy. The trouble starts when it becomes the only thing on the scoreboard. Once a team is being judged on ROAS alone, the easiest way to win is to point spend at demand that already exists: Brand search. Retargeting. Audiences that were probably going to convert anyway. The numbers might go up but the business doesn't really move. A simple example: a brand shifts budget into branded paid search and watches ROAS climb. Looks like a win. But organic clicks on the same terms drop by almost the same amount, and net new customers stay flat. Nothing was actually gained, the same revenue just got reclassified into a different bucket. This is why ROAS in isolation can hide what's happening underneath. It rewards capturing demand and tells you almost nothing about whether you're creating any. The more useful view is the one that sits next to the P&L. Contribution margin/new customer CAC. What's actually incremental versus what would have happened anyway. Those numbers are slower and messier, but they're the ones that tell you whether the business is growing or just rearranging itself.

I re-joined Darkroom as Head of AI in March - working primarily on the product side. Being a boomerang in this leg of the AI era is truly a unique experience. Perhaps unsurprisingly, the main thing that has changed around here is the speed at which we operate (and we were already moving fast to begin with). Part of that is obviously driven by the technology. A lot of value-building ideas might have previously died in a doc or a meeting somewhere. We now have the ability to cross a lot of big ideas off the list very quickly. It's easy to lose sight of the fact that projects that would have previously taken months (if they even got prioritized at all) are now taking days to complete. The work is definitely more interesting when you can shift more ideas into practice and actually start observing real outcomes. But we also move faster out of necessity. Expectations are much higher. Our competition has access to the same underlying technology that we do, and differentiating requires more energy investment than it did in the past. I think a core meta-skill that will pay dividends for organizations and individuals in this new paradigm will be the ability to find ways to preserve quality at this much more intense pace. A lot of people talk about moving to a mode of work where you're 'managing a team of AI agents', but there isn't a lot of talk about what building that skill actually looks like. That's something we're all still figuring out. But it does feel like we’re getting closer to a version of work where more time can be spent building, directly experimenting, and solving problems. 

Target just launched two new creator programs, Club Target and Target Ambassadors, as part of a bigger push into creator-led commerce. When a company at that Target's scale starts investing deeper into creator ecosystems, it usually means consumer behavior has already shifted. For a long time, influencer marketing sat outside the core business. But now, creators are becoming part of the actual commerce engine. Internally at Darkroom, we’ve been seeing this play out across TikTok Shop and social commerce for a while now. If you truly want to leverage social commerce, you need systems around: - Creators. - Performance creative. - Affiliate infrastructure. - Rapid testing. - Platform-native distribution. Creator-led commerce is now more about about building scalable systems for attention, trust, and conversion in a world where creators now sit at the center of commerce.

We launched a brand on TikTok Shop from a cold start. By month two, there were 1,400+ creator videos driving 14M impressions. - Month one did $228K. - Month two did $510K. A few years ago, the normal DTC growth timeline looked completely different. Brands would spend months buying attention, building awareness, and hoping conversion followed later. Now the loop happens much faster. - Creators post. - Distribution kicks in. - People buy directly in-feed. And perhaps the most interesting part is that the impact doesn’t stop at TikTok Shop. As visibility scaled there, we started seeing: - Demand show up across other channels. - Amazon search volume increase. - Product page traffic climb. More and more brands are starting to look at TikTok Shop differently. The reality is that it is not just another channel. It is a demand engine that can influence the broader commerce ecosystem around it.

People ask why I joined Darkroom, and the honest answer is the AI. But it's more. It's the data. Every agency right now is saying they're "AI-first," but that AI is only as powerful data underneath it. AI's value is in analyzing data, but it needs great data. Darkroom has that data. We've been running integrated, omni-channel programs for years. We've experimented as we've broken down siloes. We've watched a small change on one channel ripple into another two weeks later. We've seen audience shifts on Amazon move Meta performance in tiny shifts that become avalanches when steered into. That pattern recognition isn't something you stand up by buying an AI license. It's the accumulated result of being in the fight long enough to have seen it happen. Compared to the siloed legacy agencies, we're years ahead on what this kind of intelligence can actually do. Because the data we're training on is specific to the categories and the clients we serve. That's how we make predictions at a high rate of success, and back them up across channels most agencies still run in silos. "AI-first" is a claim everyone is making. What you train the AI on is the moat few have.

The pattern we keep seeing inside mid-market consumer brands ($10M-$200M in revenue): Paid media gets all the attention, retention gets a Klaviyo seat license and a quarterly review, CRO gets a one-off audit if budget allows, and TikTok Shop gets handed to whichever person has the most TikTok on their personal phone. Five channels, five different definitions of success. What brands need is an omni-channel approach: TikTok Shop, paid, retention, CRO, marketplaces, connects through the same growth system, the same measurement layer, and the same financial model. An example of what this could look like: → TikTok Shop rarely looks like a winner inside its own dashboard. The lift usually shows up on Amazon and the DTC site a few days later. If you're not measuring across, you'll keep underfunding the channel that's moving the rest of the funnel. Your customer moves across TikTok, Amazon, search, email, and your site in a single shopping session. Omni-channel is how consumer brands get built now.