Concierge MVP: Testing Software or a Service?
A concierge MVP tests software demand only if the manual tasks you perform for early customers can eventually be automated by code. If customers are paying for your bespoke human empathy, subjective judgment, and relationship, you are validating a service business, not a scalable product. To ensure you are building a software company, you must deliberately track your manual interventions and design tests to hand those tasks over to an automated system.
Founders often use a concierge MVP—delivering the core value proposition manually—to validate demand before writing code. While this is an excellent way to learn, it carries a hidden risk: you might accidentally build a consulting firm. This guide covers how to distinguish between founder-delivered value and repeatable product value, log your manual work, and design an automation handoff test.
The Concierge Trap: Founder Value vs. Product Value
When you act as the software, customers get a premium experience. They receive personalized attention, nuanced problem-solving, and immediate adaptations to their edge cases.
- Founder-delivered value (Service): The customer's primary benefit comes from your subjective expertise, relationship, or ability to handle undefined edge cases. Example: Acting as a fractional CFO who interprets financial anxiety.
- Repeatable product value (Software): The customer's primary benefit comes from a defined input-output process that happens to be executed by a human temporarily. Example: Taking a CSV of expenses and categorizing them into a standard P&L template.
If your customers churn the moment you replace yourself with a basic algorithm, you validated a service.
Logging the Manual Work
To prevent building a service business by accident, you must log every manual action you take to deliver value. This transforms your daily operations into a precise specification for future software.
Below is a hypothetical example of a manual work log for a concierge MVP testing an "AI meeting summarizer" concept where the founder is secretly taking notes.
The MVP Task Logging Worksheet
| Task Performed | Time Spent | Subjective Judgment Required? | Can Code Do This Today? | Next Step Toward Automation |
|---|---|---|---|---|
| Joining the Zoom call | 2 mins | No | Yes (via API) | Build bot to join calendar links |
| Identifying key action items | 15 mins | Yes (requires context) | Partially (LLM prompts) | Test standardized LLM prompt on transcript |
| Formatting the final email | 5 mins | No | Yes | Create standard email template |
| Following up on missed context | 10 mins | Yes (high empathy) | No | Eliminate this step; enforce strict inputs |
How to use this: Every week, review your log. Any task marked "Yes" under Subjective Judgment is a threat to your software model. You must either constrain the product so that judgment is no longer needed, or accept that it remains a human-in-the-loop service.
Designing the Automation Handoff Test
Once you have logged your tasks, you need to test if customers will accept an automated version of the value. This is the "Automation Handoff Test."
- Isolate one manual step: Choose a low-empathy, high-repetition task from your log (e.g., formatting the final email).
- Standardize the output: Stop customizing the output for each client. Send a rigid, templated version that looks exactly like what a basic script would generate.
- Measure the reaction: Do customers complain? Do they churn? Or do they accept the standardized output because it still solves their core problem?
If they accept the rigid, standardized version, you have proven that this specific feature can be automated. You are testing software demand.
Scoping the Transition to Code
As you successfully pass automation handoff tests, you will accumulate a list of features ready to be built. Translating these validated manual processes into a technical roadmap requires careful planning to avoid overbuilding.
For a detailed breakdown of how to translate these validated learnings into a buildable v1, read our guide on how to scope an MVP for technical feasibility. It will help you map your standardized manual tasks directly to technical requirements without bloating the scope.
Actionable Next Step: Open a spreadsheet today and log the last three tasks you performed for your concierge MVP customers. Categorize each as requiring "Subjective Judgment" or "Standardized Logic." If your entire log relies on subjective judgment, redesign your offering tomorrow to enforce stricter, standardized inputs.
Turn your manual processes into a Product Requirements Document
Idea OS evaluates your startup across market sizing, ICP, competition, and more—then generates a PRD tailored to your evaluation.
Create my PRD →New to Idea OS? Start by evaluating your idea.