Rules
Part of How Companies Can Assess Network Marketing Retention
Network marketing retention: rules to set before a trial
Network marketing retention test rules to write before day one: what counts as active, run length, group size, comparison and stop rule.
What to take away
- Test one retention change at a time, on a defined group, with a comparison basis.
- Write the formulas and the test rules before the trialwhat counts as active, run length, group size, the threshold for a result and the stop rule.
- Record the rollback condition before you touch a compensation plan, an autoship rule or a recruiting script.
- Stop the trial if the case shows pressure or sunk cost keeping a distributor in a program they want to leave.
- Treat early indicators as diagnostic, not proof.
Retention ideas reach a direct selling company through vendor decks, convention stages and distributor group chats. Few arrive with evidence that they work. A small written trial turns an idea into a result the company can read and, if needed, undo.
Write the arithmetic first
Retention is a ratio, and a ratio needs a fixed start and end. Write the three formulas below as variables before the trial begins, then substitute your own figures. Another company's numbers describe its plan, its customers and its period.
- Retention rate = distributors active at period end ÷ distributors active at period start
- Attrition cost = recruits lost in period × cost per recruit
- Payback = intervention cost ÷ (retention gain × average monthly distributor margin)
Pick one metric pair: active customer service and voluntary continuation, each read against cost and quality. A rise in continuation that comes with falling customer service is a warning, not a win. The network marketing retention page covers the wider picture.
| Figure | What it tells you | What it does not tell you |
|---|---|---|
| Retention rate | Who was still active at the end compared with the start | Whether they sold to customers or wanted to stay |
| Attrition cost | What the recruits lost in the period had cost | Why each person left |
| Payback | How long an intervention takes to earn back its cost | Whether the gain would have happened anyway |
Fix the test rules before day one
A trial is readable only if its rules exist before the data does. Write each rule below on the trial record, with a date and an owner, before the first distributor receives the change.
| Rule | What to write down | Why it matters |
|---|---|---|
| Definition of active | The exact action that counts a distributor as active, applied the same way at period start and end | A definition that moves mid-trial changes the rate without changing a single distributor |
| Run length | A start date, an end date and a review date | An early read can show a gain that later disappears |
| Group size | How many distributors are in the trial group and in the comparison group | A group so small that one exit moves the rate cannot support a conclusion |
| Comparison basis | A similar group on the current rules, read over the same period | Without it, a seasonal rise looks like a result |
| Threshold for a result | The gap between the two groups you will accept as real, set before you see the data | A threshold chosen afterwards fits whatever happened |
| Stop rule | The conditions that end the trial early | A gain that traps a distributor who wants to leave is not a gain |
Read the result as a difference. If the trial group's retention rate rose and the comparison group's rose too, only the gap between the two changes belongs to your intervention. Report that gap beside the cost, and state what the trial could not show.
Decide what the change touches
Name the workflow before the group: a compensation plan, an autoship rule or a recruiting script. Each has its own rollback condition, the result that tells you to undo the change, and it must be written before anyone touches production work.
Then check where the idea stands. A platform preview shown at a convention is not a shipped change, and an announced autoship rule is not in force. Only shipped changes, or a limited release tested inside its eligible group, belong in a trial. An announcement is a plan, and an inference is a research question.
Watchlist Status Decides Action
Shipped
- Evidence
- Primary notice
- Effective date
- Passed
- You may
- Change production
- Rollback
- Written first
Limited release
- Evidence
- Restricted eligibility
- Effective date
- Varies
- You may
- Test eligible group
- Rollback
- Within group
Announced
- Evidence
- Future date
- Effective date
- Future
- You may
- Plan only
- Rollback
- Not yet
Inferred
- Evidence
- Adjacent evidence
- Effective date
- None
- You may
- Research question
- Rollback
- Not yet
Name the owner who reads the primary source, and save that source with the date you read it. An item without an owner is a rumor with a date on it.
Keep the compliance boundary visible
Retention work touches income claims, compensation disclosures, recruiting language, consent and records. The FTC Income Disclosure Guide and Section 5 of the FTC Act govern what a company may say about earnings.
The FTC's Business Guidance Concerning Multi-Level Marketing applies too, and the Endorsement Guides at 16 CFR Part 255 add the rules for testimonials and disclosures.
State attorney general rules vary from state to state, and so do return windows and exit rights. Keep the relevant state cooling-off laws in the file next to the change. Confirm the current requirement with your state regulator or a licensed attorney before a script, a testimonial or a landing page goes live.
Compliance File And Vendor Checks
- Approved scope
- Exclusions
- Version
- Owner
- Decision date
- Vendor development practice
- Supply-chain exposure
- Vulnerability management
The working file should show approved scope, exclusions, version, owner and decision date. When a vendor pitches a retention tool, the CISA software acquisition fact sheet points buyers toward development practice, supply-chain exposure, deployment and vulnerability management. Those are purchase-record checks, not an approval. Your team still owns the facts and the decision.
Run the trial on one defined group
A retention trial needs a defined population and a comparison basis. Give the change to one group, keep a similar group on the current rules, and read both over the same period against cost and quality.
Watch one normal path and one failure path, where a distributor leaves despite the intervention. The failure path shows whether exit stays easy, and whether anyone is applying pressure to hold a person in the program.
The NIST handbook says the choice of an experimental design depends on the objectives of the experiment and the number of factors to be investigated. It does not answer your local question. The practical lesson is to change one thing at a time, so that any difference has one candidate cause.
GAO's guide to evaluations covers the process of designing evaluation studies and the key issues in study quality. Use that discipline, keep the federal-program context in view, and avoid causal claims your method cannot support. Record the source date and limits beside the decision.
The trial in six steps
- Write the three formulas and the test rules, and choose the metric pair.
- Name the workflow, the owner and the rollback condition.
- Confirm the status of the change and the date of the primary source.
- Pick the trial group and the comparison basis.
- Watch one normal path and one failure path.
- Compare against cost and quality, state what is unknown, and set the next review.
Stop the trial if the case shows pressure or sunk cost delaying a sound exit. A retention gain that traps a distributor who wants to leave is not a gain the company can defend.
Expansion depends on a result that reproduces with the people, rights, capacity, systems and budget you actually have. Read it against regional markets and retention before copying it to another market, and build the next plan by developing strategy from evidence, not habit.
Common questions
What is the first decision in a retention trial?
Name the owner, the audience and the outcome before reading any source. Then write the evidence that would stop or change the work. A review without a stop condition becomes a justification for whatever the company already planned.
Can the trial use another company's retention figures?
No. Substitute your own figures in the formulas. Another company's numbers describe its own plan, customers and period, so they cannot tell you whether your change worked.
What if an error reaches a public page or recruiting script?
Correct it there and keep the record of what was said. A fix made only in private coaching leaves the public statement standing, and the record shows what changed and when.
Does a retention tool decide the outcome?
No. Software supports a decision the company still owns. Confirm the compliance requirement with your state regulator or a licensed attorney, and confirm the tax treatment of distributor payments with the IRS or a CPA.
