Two products, two user bases, one backend

When we first launched One Final Message, the product had one signup flow, one dashboard, one pricing page, and two kinds of users showing up in it. That was the problem. This post is about what we learned from watching that play out, and what we are changing in response.

The two user bases

We can describe the two users without making it about personality.

The planner.

The planner has time. They are organizing their digital life because something prompted them to think about it - turning fifty, settling an estate for a parent, finishing a will with an attorney - and they are comparing options. They want to add multiple messages, multiple recipients, supporting documents. They want to come back and refine over months. Their decision is rational. The buying surface they expect looks like a planning tool: tiered plans, feature comparison, a long onboarding that they don't mind because they're not in a hurry.

The person with a trigger.

This user is not browsing. There is a specific event: a court appearance, a deployment, a hospital admission, a protest they are about to attend, a diagnosis received this week. They want one message held for one person, scheduled to send if they stop responding to check-ins. They want the setup done in five minutes. They are not interested in features they will not use. They will likely never log back in to add things; the value of the product to them is that it is set and forgotten until needed.

These users have almost nothing in common with each other except a fundamental need: a message that gets to the right person at the right time. That is enough to share infrastructure. It is not enough to share a product.

Why one product could not do both jobs

We tried, for about a year, to make it work as a single product with a smaller tier (The Lockbox) carved out for the urgent user. The architecture was clean enough. The user experience contradicted itself.

The dashboard a planner needs is a workspace: vitals, assets, a readiness ring, prompts to add the next thing. The dashboard a triggered user needs is a confirmation screen: yes, your message is set, here is when we will check on you next, here is how to pause. The same screen cannot do both. We tried having one screen do both, and it ended up nudging the triggered user toward features they did not want and would not buy, while leaving the planner feeling like the experience was thinner than the product they had paid for.

The marketing had the same contradiction. Copy that sells the urgency of "what if something happens" lands wrong when the buyer is methodically planning their estate. Copy that sells the calm of "organize your legacy" lands wrong when the buyer is about to enter a courtroom. Trying to write one set of words for both audiences produced copy that did neither job well.

The partner-channel problem was the one that forced our hand. We started talking to immigration legal aid clinics about referring users to the urgent product. The clinics correctly asked whether a referral would expose their clients to upsell. We could not say no, because the existing product was structured as a funnel. We could not redesign the funnel without compromising the planning product's revenue model. We had a structural conflict between the two user bases that we could not resolve inside one product.

What we changed

We split the products.

One Final Message remains the planning product. It has been restructured into The Lockbox, The Safe, and The Vault plans aimed at long-term planners. Marketing, partnerships, and copy on this side are oriented toward legacy planners, estate attorneys, fiduciaries, and financial advisors. The buying decision is logical. The dashboard is a workspace. The cadence of the user's relationship with the product is measured in months and years.

Beacon is the new product for the triggered user. It lives at milkweed.foundation, operated under the Milkweed Foundation. The setup is three steps: recipient, message, schedule. The dashboard is a single confirmation screen with a countdown. There is no upsell anywhere inside the activated product. The buying decision is emotional and time-pressured. The cadence of the user's relationship with the product is daily check-ins, indefinitely, with no second product to be sold into.

Both products share the same Cloudflare Workers backend, the same Durable Object data layer, the same authentication, and the same delivery pipeline. The shared work has already been done. The split is at the brand and product-surface layer, not the architecture layer.

What this enables, specifically

The split is not cosmetic. It changes what we can do. We can now talk to public benefit organizations without the upsell shadow. Immigration legal aid clinics, hospice services, protest legal observers, and similar groups can refer clients to Beacon knowing that the product itself does not push referrals toward an unrelated subscription. The Milkweed Foundation framing is institutional, not commercial, and the absence of upsell is enforced in the product, not in policy. A partner does not have to trust our intentions; they can look at the product and verify.

We can now talk to fiduciary, estate-planning, and legacy partners without the urgency overhang. Attorneys, financial advisors, and estate planners can refer clients to One Final Message knowing that the experience they get is the planning product, not a service that pivots into a daily-check-in cadence aimed at people in crisis. The marketing tone on this side stays calm, methodical, and competent - appropriate to the buying context.

We can now write separate emotional registers. Calm and methodical for the planner. Plainspoken and protective for the triggered user. We were trying to do both before, and the result averaged out to something that landed cleanly with neither group.

We can now price the two products independently. Planning customers price the value of long-term organization. Triggered customers price the value of a single message held reliably. These are different pricing problems; the previous one-product structure forced them into the same pricing logic.

What this means for current customers

Existing Beacon-plan customers on OFM will be migrated to Lockbox plans with their current schedule, recipient, message, and pricing preserved. Their existing pricing is grandfathered indefinitely. 

No customer of either product is asked to choose between products, and no customer is silently downgraded. Customers who decide they want the other product can switch with a one-click action in account settings; that capability exists in both directions.

What this does not change

The underlying delivery commitment is unchanged. A scheduled message goes to its recipient at the scheduled time, regardless of which product the schedule was created in. The infrastructure that handles that delivery is shared. The security and privacy posture is unchanged. Both products use the same authentication, the same encryption-at-rest, the same audit logging, and the same recipient-token model for delivery confirmation. A split at the brand layer does not create new privacy boundaries; the existing boundaries continue to apply.

The team is unchanged. We are not hiring a separate Beacon team. The same engineers maintain both products. The brand and partner-channel work for each product is divided between us as founders.

What is next

Beacon is live now. The Milkweed Foundation is being set up as the institutional sponsor; details on its legal structure will follow when they are finalized. Partner conversations on the Beacon side are already underway and we expect to expand that list before launch. Partner conversations on the One Final Message side are continuing with the legacy-planning audience we have been talking to throughout.

If you represent an organization that works with people in either of these positions - long-term planners on one side, people in immediate need on the other - we are interested in talking. There is no upsell between the two products, in either direction, by design.