Network marketing retention benchmarks: definitions, windows, and how to recalculate

What to take away

The numbers, and what each one actually measures

Most "retention benchmark" claims in direct selling are one of four measures wearing the same word. Name the measure before you quote a number.

Worked example: a company reports a 12-month retention rate from a starting cohort of 4,000 distributors, with 1,400 still ordering. That reads as 35%. Exclude 600 enrollees who never placed an order and the base falls to 3,400, so the same 1,400 now reads as 41%. Same distributors, same behavior, different headline.

The table is illustrative. Treat each range as a starting point until you have the source document in hand. The nearest named source for the disclosure expectations behind the ranges is the FTC's Business Guidance Concerning Multi-Level Marketing, published in January 2018. No single industry body publishes one audited retention rate for the channel.

Four things sit behind those numbers. The activity window, at three, six or twelve months, decides how long one order keeps a distributor in the count. A minimum qualifying volume decides whether a small order counts at all.

A rank filter cuts the base down to its top slice. Cohort maturity decides whether you are reading a group that has had twelve full months or a rolling snapshot.

The FTC's business guidance on multi-level marketing and the DSA's member guidance both push companies toward disclosing earnings and attrition honestly. Neither publishes a retention benchmark.

The DSA's statistical reporting covers sales and seller counts instead, with recent U.S. retail sales in the $35 billion to $40 billion range. That gap is why so many figures circulate without a denominator attached.

The FTC's 2016 settlement with Herbalife carried $200 million in consumer redress and a restructuring of the U.S. business. It required retail customers and distributors to be tracked separately. What changed was Herbalife's own reporting, not the channel's.

Why the same rate reads differently at two companies

A 35% twelve-month retention rate sounds strong until you learn the company counts anyone who placed a single order, including a $40 starter kit. A 22% rate looks weak until you learn it counts only distributors who hit a rank.

Three factors move the number more than any retention program does:

Change any one and the rate moves several points without a single distributor behaving differently.

Window length alone does much of the work. Nu Skin's reported customer counts use a three-month purchase window, and USANA's active associate definition uses twelve months, so identical buying behavior produces different headline rates at the two companies.

Operating rules change the base as well. The 1979 FTC decision on Amway turned on rules like these: a 70% rule requiring that at least 70% of a distributor's purchases be resold, a 10-retail-customer rule, and a buyback policy for unsold inventory.

A base held together by retail and buyback rules churns at a different rate from one held together by autoship alone.

Recalculating a figure before you use it

Take a published rate and rebuild it. Suppose a company reports a 12-month retention rate of R, from a starting cohort of N distributors and a retained count of K. Then R = K / N. Ask for N and K separately.

Run the same test on the window. A distributor who places one order in March stays inside a twelve-month count until the following February, and inside a three-month count only until the end of May. One order, two retention stories.

Do this arithmetic with your own figures before you adopt anyone else's threshold. Where the inputs are not published, say so in your file and use your own cohort instead.

Checking the source before you cite it

A benchmark is a claim, and claims need a paper trail. In direct selling that trail runs from the company's income disclosure statement to the DSA's member reporting and on to the FTC's business guidance on multi-level marketing. The W3C guidance on information and relationships is a useful discipline here.

A table of rates must carry its structure in the markup. It must not carry that structure only in the visual layout. The relationship between a rate and its population must survive being copied out.

The CISA software acquisition fact sheet applies the same logic to tools: know what you bought, from whom, and what it exposes. If your retention dashboard comes from a vendor, that record belongs next to the benchmark.

The GAO evaluation design guide sets out how to link an evaluation question to the evidence that can actually answer it. Borrow the discipline, not the federal context, and do not claim causation your cohort cannot support.

For privacy handling of distributor records used in the analysis, the NIST Privacy Framework starting guide describes a voluntary process for identifying risk and assigning owners. It is a management aid, not legal clearance.

A worked benchmark card

Run one card per figure. Two cards on the same page usually reveal that one number is doing work the other contradicts.

Common questions

Why do retention benchmarks differ so much between sources?

Because "retention" is not one measure. A twelve-month distributor rate, a monthly active rate and a 90-day autoship persistence rate all get published under the same label, and each uses a different denominator.

Can I use an industry retention benchmark as a target?

Only after you rebuild it from its own inputs and confirm the population matches yours. If the source does not publish N and K, use your own cohort and treat the published figure as context, not a goal.

What should I do when a benchmark has no published method?

Record the gap, note the date you checked, and set your own threshold from your own data. For the terms behind those thresholds, Common retention questions covers the definitions to settle first, and the retention overview explains how the measures fit together.