May 1, 2026 | The Product Manager’s Journal
The Take
I built a persona tester this week. It changed how I think about feature decisions.
Here’s the setup. I’ve been wrestling with a packaging copy decision for Trevean for two weeks. The kind of small choice that should take an hour, but somehow eats six. Do we label the back of the jar with the farmer’s name or the country of origin? Do we put the harvest date in the same font weight as the lot number, or step it down? Does “tap your phone here” sit above or below the NFC mark?
Every decision had a defensible answer. None had the right answer. And I noticed something embarrassing about my process: I was making these calls based on a founder who lives inside this product, and would respond to. Not what an actual customer would.
So I built a persona tester with Claude.
Quick description of what it is. I wrote up four detailed personas based on the customer interviews Rushi and I have done over the past nine months. The Health-Forward Mom in suburban Chicago. The Spice-Curious Apartment Cook in Brooklyn. The Retired Foodie who travels to Italy twice a year. The Restaurant Sous Chef, who shops for himself on his off days. Each persona has a budget, a kitchen, a story about why they care about food, the questions they actually ask in the aisle, the words they use, and the things that turn them off.
Then I gave Claude a system prompt: respond as exactly one of these personas, in their voice, with their concerns, their vocabulary. Don’t summarize. Don’t hedge. Don’t break character. If something is confusing, say it’s confusing. If something feels like marketing, call it out as marketing.
I dropped my packaging copy in. I asked each persona, one by one, what they noticed first, what they didn’t understand, what made them want to put the jar back, and what would make them reach for it.
The Sous Chef told me he didn’t care about the farmer’s name and would be annoyed if it crowded out the harvest date. The Apartment Cook said the harvest date felt clinical and asked why I wasn’t telling her how actually to use the spice. The Health-Forward Mom asked if “tap your phone” meant she was being tracked. The Retired Foodie wanted to know more about the farmer than I’d written and asked where the farm was on a map.
Four reactions. Four different problems. None of them was the problem I thought I was solving.
A few things to be clear about, because I know how this gets misread:
This is not a replacement for talking to actual humans. Real customer interviews are still the only way to discover problems you didn’t know existed. The persona tester only knows what I taught it. It can’t surprise me with a need I never put in.
This is a sharper tool than a feature spreadsheet. When I’m choosing between two defensible options, the persona tester is faster than a survey, more honest than my own gut, and stops me from optimizing for the version of the customer who looks the most like me.
The persona quality is everything. A vague persona produces vague feedback. A specific one, with quotes from real interviews, the specific language she uses, and the specific reason she stopped buying her last brand, produces feedback I can act on within an hour.
The PM lesson buried in this is the part I keep coming back to. Most product decisions don’t fail because we picked the wrong answer. They fail because we picked an answer that worked for one customer and assumed it would work for the rest. The persona tester forces me to hear from four customers in my head before I commit, and four wrong answers in advance is cheaper than one wrong answer at scale.
The full write-up is live. It walks through the prompt structure, the persona template, and the exact sprint that this saved me. If you’re choosing between two defensible options on anything this week, this is the tool I’d point you at first.
→ Read the full post: How to Build a Persona Feature Tester With Claude (and Why It Saved Me a Sprint)
Spice Route Signal
The synthetic dye reckoning is finally hitting the spice aisle.
I want to flag a regulatory shift that hasn’t gotten enough attention from food product folks: the FDA’s continuing tightening of synthetic food colorings. Red No. 3 was banned for food use in early 2025 with a January 2027 compliance date, and the momentum hasn’t stopped there. State-level bans on Red No. 40, Yellow No. 5, and Yellow No. 6 are stacking up, retailer phase-outs are accelerating, and the entire “clean label” conversation is moving from a marketing posture to a compliance requirement.
Why this matters for spice products specifically: a surprising amount of what’s labeled “paprika,” “chili powder,” or “saffron” on a grocery shelf is color-corrected with synthetic dye to hit a visual grade. Producers do it because buyers grade by color, and a slightly pale spice gets discounted. The pigment is cheap. The labeling rules have historically allowed it under “spices” or “color added.” Most consumers have no idea.
When the regulatory floor on synthetic dyes rises, it doesn’t just hit cereal and candy. It hits the spice supply chain in ways most CPG teams haven’t priced in yet. The certificates of analysis enterprise buyers already require will become table stakes for retail. The producers who’ve been color-correcting will need new processes. The brands that built their entire identity on “what you see is what was grown” will look prescient.
This is the same dynamic I keep writing about: regulatory tailwinds favor the brands that were already doing the harder thing. If you’re a PM in food, ingredients, supplements, or any category where color or appearance affects perceived quality, this is a 2027 problem you should be looking at in 2026. The companies that get ahead of it will own the shelf when it lands. The companies that don’t will spend the back half of 2027 in a compliance scramble.
I’ll keep tracking this and report back as the state-by-state map gets clearer.
From the Trenches
Lid update, since a few of you have actually started asking about it. Yes, I see the irony.
We confirmed it. The cork liner does insulate the NFC tag enough to break clean phone reads. We scraped the cork off a test lid, applied the tag directly to the bottom of the lid, and the reads came back perfect. So the architecture I described last week, NFC tag adhered to the lid, cork liner stacked underneath as a separate component, is the path. The Chinese suppliers, all five of them I reached out to, are on it. Responsive. Sending videos of the assembly process. Samples in transit if I need them.
My US supplier? Crickets. Two weeks now, no reply, no acknowledgment, no nothing. I keep checking my spam folder, like the email is somehow my fault. It isn’t. The math keeps getting clearer.
Now for the part where I confess. I keep coming back to this lid design. I know I shouldn’t. The decision is functionally made. The samples are coming. The architecture is sound. And yet, every time I drop a render of it into a slide, I get those oohs and aahs from anyone who sees it, and the reptilian founder part of my brain whispers what if there’s still one more variation worth exploring?
There isn’t. There rarely is. This is exactly the kind of late-stage tinkering I wrote about in the Design Freeze post a few weeks back, and apparently writing about it is not the same as practicing it. I’ve been thinking I should run it by one of the personas I will now call Busy Beth for a quick chat, mostly so a clear-eyed person can tell me what I already know. And now that I’ve committed this in writing, I’m sure to hear the screams from across the office to stop with the lid design.
For the record: I will stop. Today. Writing it down here is the commitment device. If you see me posting another lid render next week, you have permission to reply with a single word. Beth.
Parallel thread: the Claude Designer website work is still going. The first impression continues to hold. I’ll have something tangible to show you by Issue #8.
From the Rack
Turmeric, part two. Last week I wrote about how it’s fake. This week, why it matters that it isn’t.
I wrote last week about lead chromate adulteration in turmeric and the PM lesson about surface metrics that get gamed. The piece was important to write. It was also, in hindsight, a downer. So this week I want to flip the lens.
At the molecular level, turmeric is one of the most studied and validated spices in the world. The active compound, curcumin, is the subject of more peer-reviewed research than almost any other plant-derived bioactive. Anti-inflammatory effects. Antioxidant capacity. Emerging evidence on cognitive function, joint health, and metabolic markers. None of this is folklore. It’s clinical literature, growing every year.
What’s striking is that human cultures figured this out three thousand years before we had the science to explain it. Indian, Southeast Asian, and Persian cuisines have used turmeric daily, often with black pepper to enhance absorption, often with fat to stabilize curcumin, and often in combinations that match exactly what modern bioavailability research now confirms. The traditional cuisine got there first. The lab spent two decades catching up.
Here’s what I keep thinking about. The reason turmeric is fake is the same reason it’s worth faking. It’s valuable because it does something real. The molecule has been doing the work for millennia. The adulteration problem isn’t a story about turmeric. It’s a story about a market that lost track of what it was actually buying.
The PM angle: the products worth protecting are the products that do something real. If your product is so trivial that fakes don’t bother with it, that’s not a competitive moat. That’s a signal. Build the thing that’s worth faking. Then build the verification layer that makes faking it pointless.
→ A friend sent me a great, longer read on turmeric’s broader cultural and health significance this week. If you want the link, hit reply and I’ll forward it.
On My Desk
“Super Body” by William Davis.
Not the typical business book. I picked it up because a customer recommended it, and I keep my customer recommendations on a separate shelf. Davis is a cardiologist who writes about the gap between the way the modern food system is structured and the way the human body actually works. The argument, stripped of the marketing copy that usually wraps this kind of book, is straightforward: the further a food is from its unadulterated, unprocessed form, the worse it tends to be for the system designed to digest it.
I’m not a doctor. I’m not going to summarize the medical case. What I will say is that reading Davis felt unsettlingly familiar for someone running a spice company. The principles he articulates about food, that processing strips out things we don’t yet understand, that supply chains insert ingredients consumers never agreed to, that “natural” on a label is doing a lot of unaudited work, are the same principles Trevean is built on, just applied in a different idiom.
The thing that stuck with me most was a small observation buried halfway through the book. Davis points out that traditional cuisines, those that evolved over centuries in specific places, rarely include the ingredients most reliably associated with modern chronic disease. Industrial seed oils. Ultra-processed grain products. Synthetic additives. The cuisines that humans actually evolved alongside, the ones that get expressed in spice blends from specific regions, used real ingredients in real ratios because the alternatives didn’t exist yet.
Trevean’s whole product thesis, that a blend assembled with intent from real, traceable, unadulterated ingredients is functionally different from the same ingredients separated and recombined from a commodity supply chain, sits inside this argument cleanly. I didn’t expect that going in. About 320 pages. Read it on a long weekend.
That’s the Rack
Thanks for reading Issue #6. I’m Dan Blizinski, founder of Trevean Spice and the person behind The Product Manager’s Journal, where I write about PM frameworks that come from actually building things, not just theorizing about them.
New here? Grab the free Startup PM Toolkit. Five frameworks I actually use, not just talk about.
Question for you this week: what’s a decision you’ve been over-tinkering on past the point of useful return? Your version of my lid? Hit reply. I read every one. And if you’ve built a persona tester of your own, I want to see it.
The PM’s Spice Rack is published weekly on The Product Manager’s Journal and on LinkedIn. Subscribe to get it in your inbox.