BlueAlpha

Waymo is in 11 cities, targeting 1M rides a week, and just raised $16B. Their product is everywhere. Their paid media footprint is a fraction of their competitors. Let’s get into why. I ran Waymo through my team’s competitive benchmark system against Uber, Lyft, and Zoox. Overall: 54/100. Second in the peer set. Here’s the breakdown: 1. Their Meta program is building a brand while competitors fight on pricing. > 79/100 Meta health score, best in the peer set > 53% brand story and use case while Uber and Lyft lead with promos > Formalizing creative retirement would let the program build more cleanly as it scales into new markets 2. Google is more actively managed than any competitor, and the opportunity is to pair that with a foundation. > 0% evergreen dependency + 35% recent launch rate. Super hands-on compared to peers > 93% video is a bold outlier in a category that runs 92-100% text > A small evergreen core alongside the active cadence, plus text and image variants, would significantly reduce structural risk 3. The creative library is the biggest lever as Waymo enters 20 new markets this year. > 1,234 total estimated ads vs Uber at 55,738 (45x gap) > More creative surface area means faster learning by city, audience, and use case, necessary for scale at this pace A brand scaling this fast across this many new markets needs paid media that scales with it. The Meta program is already there. Google and creative volume are where the next chapter gets written. Built using public data. Comment "benchmark" and I'll run one for you.

We built BlueAlpha for marketing teams who are done taking their ad platform’s word for it.   Last quarter, a consumer fintech scaling iOS acquisition came to us with a problem we see all the time: 30-60% of their Meta spend was getting credited for conversions it didn't cause. Platform metrics said keep spending. The P&L told a different story. Our incrementality tests flagged $2M to reallocate. They moved it. Pipeline held. CAC fell by 40%. This isn't a one-off. It's a pattern we see constantly, and the reason is simple: the tools measuring your spend have a direct interest in how you allocate it next. Three things worth checking this week if you're running a paid media program:   → What percentage of your “conversions” survive a holdout test? → Can you answer "what happens if we turn off Channel X?" with data, or just a guess? → When did you last prove a channel was incremental rather than assume it?   If you want the framework we use to answer these, follow BlueAlpha. We publish the methodology, not just the results.

The BlueAlpha platform is getting an upgrade. New look, new feel, and a big shift in how you interact with your data. The old version felt fixed. Useful, but you were limited to what had already been configured. The update we’re shipping is way more flexible: ask for one number, one chart, a specific breakdown you've never pulled before, and get it on the spot. Pre-fabricated dashboards are still there for the things you check every day. But we kept hearing the same thing from clients: the most valuable moments in the product were when they found something they weren't looking for. The new design is built around that. More on what it looks like once it's fully live.

MCPs have completely changed how our team operates. Here's one example that I think is underrated. We’re making our MMM queryable in plain English. Anyone on the team can ask it a question and get an answer in seconds. No data ticket, no notebook, no waiting. The problem we kept running into before this setup: You spend months building the model, the priors are calibrated, everything lines up… and then it goes into a shared drive and no one knows how to interrogate it. The data team becomes a bottleneck. Marketers fall back on platform ROAS and gut instinct. The model never gets used the way it should. The fix was connecting our fitted Meridian model to Claude via an open-source MCP server. From there, anyone can ask Claude directly: > "Which channels have headroom for more spend and which are saturated?" > "What happens to conversions if we cut the budget by 15%?" > "Show me the last six weeks of contribution trends. Are we seeing diminishing returns anywhere?" Each question triggers the right model query behind the scenes. The answer comes back in seconds, grounded in the model real-time, not a stale drive from last quarter. Curious to hear what other teams are doing.

The more teams I meet, the clearer this gets: everyone is optimizing spend, almost nobody is tracking creative fatigue. MMMs aren't designed to catch it, so even the smartest teams overlook this really simple fix. We just shipped an update to how BlueAlpha detects creative fatigue, and it’s something that sounds niche but protects massive budget. When a channel starts underperforming, most measurement tools tell you to cut spend. But there are at least three different reasons a channel goes bad, and spend is just one potential reason. Your campaign settings might be misconfigured. Or your creative has gone stale. Your ads have run long enough that your audience has tuned them out. Reallocating budget doesn’t solve that problem. I've seen teams pull spend from a channel that was actually working fine. The creative was just exhausted. The fix was a refresh, not a reallocation. That's one piece of the puzzle we’re helping teams solve for. When a bad signal comes through, our creative fatigue agents analyze what's actually driving it. Spend-related, campaign configuration, or creative exhaustion. Most of the time, teams never find out. They see a bad signal and cut spend. Sometimes that's right, but often it isn't.

The first thing you see when you open BlueAlpha just got a lot more useful. Every time a client opens the product, we want that first thing they see to answer one question: what should I do today? We just got a lot closer to that. Here's what changed: 1. Ad platform data and MMM data now live in the same view. Your live channel performance alongside what the model is actually telling you to do about it, no more toggling between the two to connect the dots. 2. The Strategy Agent is now at the forefront. Instead of digging for it, you can riff on recommendations the moment you open the product. What's working, what to cut, what to shift, right there, first thing. 3. The two in combination give you a live feed of your marketing performance with the actionability layer built in. You're not just seeing your data, you're seeing your data with a point of view on what to do next, every time you log in. The thinking behind it: the marketing teams getting the most out of BlueAlpha aren't opening it to review a report. They're opening it to make a call. The overview page needed to match that workflow, live data and a clear next step, without having to go looking for either.

You can now explore BlueAlpha without getting on a call with us. We've always believed you shouldn't have to trust a sales pitch to understand what you're buying. The product should speak for itself. Until now, the only way to actually see it was to get us on a call first. That's been a friction point we wanted to fix. So, we’ve launched a tailored interactive demo. Click through the full platform, see exactly what it does, no booking required. Here's how it works: you land on the demo, select your industry  and you're in. The experience is tailored to your category so what you're seeing is relevant to how the product would look with your data. Try it out for yourself - we’d love to hear what you think. https://bluealpha.ai/demo

The way B2C and D2C brands compete for customers is changing faster than most marketing teams realize. I spent the last few weeks writing out our thesis on where consumer marketing is headed, and I’m sharing a three part breakdown this week. For the last decade, the game was simple: show up when someone searches. Win the keyword, win the click, win the customer. SEO gets you there organically. Performance marketing buys your way there. Both work. Both are also entirely reactive: you're competing for a moment of intent that someone else initiated. That model is breaking down. AI agents are starting to make purchasing decisions on behalf of consumers. And as that shift accelerates, the moment of decision moves from the point of purchase to long before it. The agent doesn't wait for the consumer to search. It already knows what they prefer. It's absorbed thousands of signals (content, reviews, brand presence, community discussions) and formed a view before any transaction happens. The purchase becomes the execution of a preference already formed. Which means the marketing lever that actually matters is no longer the one that captures intent. It's the one that shapes preference. Most marketing infrastructure isn't built for that. Most measurement systems can't even see it. More on what that actually means for your marketing team this week. Full piece linked in the comments if you want to read ahead.