I think every growth loop has a bill.
Sometimes the product pays it through incentives, infrastructure, or support. Sometimes the sender pays it by risking a little social credibility. Often the recipient pays it through interruption, evaluation work, unwanted disclosure, or one more account they never asked to create.
The dashboard usually records the invitation and the conversion. It rarely records who absorbed the cost between them.
That omission can make an extractive loop look efficient.
The cost can sit outside the actor
Economics gives us the language of externalities. A choice creates a cost or benefit for someone who was not fully represented in the decision. The CORE Econ treatment of external effects uses the concept for larger economic problems. A product invitation is much smaller, but the analogy is useful.
The sender gets the storage credit, unlocked feature, or easier collaboration. The product gets another reachable address. The recipient gets a message, a decision, and perhaps a setup task.
Those costs can be reasonable. Collaboration requires coordination, and a well-chosen invitation can be genuinely helpful. The problem begins when the loop optimizes the sender’s reward while making the recipient’s cost invisible.
Imagine a hypothetical photo organization product. A user receives extra storage for inviting five family members. The product preselects contacts, drafts urgent copy, and sends repeated reminders. The sender gets value immediately. Each recipient has to work out what was shared, whether an account is required, who can see their photos, and how to decline.
The loop may convert. It may also spend family trust the product did not earn.
Reputation is shared infrastructure
Email systems make the externality more concrete. Google’s email sender guidelines set requirements for authentication and spam rates, with one-click unsubscribe required for marketing and subscription messages from bulk senders. These rules exist at the channel level, but they expose a product truth. Recipient reactions influence whether future messages are trusted and delivered.
One aggressive loop can damage more than its own campaign. Complaints and disengagement can weaken the sender reputation used by receipts, security alerts, and truly useful collaboration messages.
The FTC’s CAN-SPAM compliance guide describes legal requirements for commercial email, including accurate headers and a working opt-out. Compliance is the floor, not a complete product standard. A message can be lawful and still transfer too much work or embarrassment to its recipient.
Sender reputation also has a human version. When a product writes as if a friend personally vouched for something they barely saw, it borrows that friend’s name. If the message is irrelevant or manipulative, the cost lands partly on the person whose identity made it persuasive.
Count more than sends
I would inspect a loop across four kinds of cost.
- Attention cost covers the interruption and time required to understand the ask.
- Decision cost covers evaluating fit, permissions, price, and safety.
- Social cost covers the sender’s risk of seeming careless, promotional, or intrusive.
- System cost covers complaints, blocks, support, and damage to channel reputation.
The costs vary by relationship. A teammate expecting access to a shared project is not the same as an address-book contact receiving a generic referral. A co-parent asked to approve a school form is not the same as a cousin pulled into a storage promotion.
Context changes whether the invitation is service or tax.
The artifact I want is a loop cost map
Map one full cycle from the triggering action to the recipient’s resolution.
| Step | Value received | Cost imposed | Cost bearer | Evidence |
|---|---|---|---|---|
| Sender selects contacts | Storage credit | Contact review and social risk | Sender | Abandonment before send |
| Recipient gets email | Awareness of shared album | Interruption and ambiguity | Recipient | Opens, complaints, research |
| Recipient previews | Can judge relevance | Privacy and permission review | Recipient | Preview exits, support topics |
| Recipient joins or declines | Access or closure | Setup or refusal effort | Recipient | Completion and decline rates |
| Product follows up | Possible reminder | Repeated attention and reputation | Recipient, sender, product | Unsubscribes, spam reports |
Add what the recipient can understand without creating an account, whether the sender approves the exact message, and where a clean decline ends the loop.
My hypothesis might be the following.
If recipients can see who invited them, what object was shared, and what access joining grants before signup, more accepted invitations will reach a meaningful shared action while complaints and sender regret do not rise.
That test is not trying to maximize acceptance alone. It asks whether better-informed participation produces a healthier loop.
I would also interview senders whose invitations were ignored. Not to teach them better tactics, but to learn whether they understood what the product sent in their name.
A sustainable loop leaves both sides better off
Growth teams often describe loops as self-reinforcing. That phrase can hide the direction of reinforcement.
A good loop makes the product more useful to the current user and creates a fair opportunity for the next person. A bad loop rewards the current user by spending the next person’s attention and the channel’s reputation.
The difference will not appear in a viral coefficient by itself.
Before scaling an invitation mechanic, I would map the value and cost for every participant, especially the person who did not initiate the action. If the economics work only because their effort, uncertainty, or annoyance is free in our model, the model is incomplete.
The recipient is not empty space between a send and a conversion. They are a participant in the loop, and they receive the bill whether we measure it or not.
