Designing a Lead Return Policy Both Sides Will Sign

Cedar Line Leads and an auto buyer they had worked with for two years stopped working together over eleven dollars. Not really, of course. They stopped over a return dispute where the buyer sent back 14 percent of a week's volume citing "bad quality," Cedar Line refused the credit, and both sides discovered their insurance lead return policy was a single sentence in an email from 2024. Illustrative names, ordinary ending. The contract clause that gets the least attention during onboarding is the one that kills the most partnerships, and it kills them at the worst possible moment: when volume is finally good.

Why the return clause breaks partnerships

Because it is asymmetric in a way nobody says out loud. The buyer holds the money and the data. The seller holds neither. A buyer who returns aggressively can effectively reprice a deal after the fact, and a seller who refuses every return can effectively sell garbage with impunity. Both sides know this, both sides assume the other might do it, and an undefined insurance lead return policy leaves that suspicion sitting in the middle of the relationship doing quiet damage.

The fix is not trust. It is specificity. When the reason codes are enumerated, the evidence standard is stated, and the cap is a number, there is nothing left to litigate by feel.

The four parts of a workable lead return policy

Reason codes that survive an argument

A good code is objectively checkable by the seller without needing the buyer's word for it. Test every proposed code against that sentence. Cedar Line's rebuilt insurance lead return policy landed on six.

That last row does the heavy lifting. Non-interest is a conversion outcome, not a defect, and pricing already accounts for it. If a buyer needs relief on non-interest, the honest instrument is a lower CPL or a tighter filter, not a return. Sellers who accept "not interested" returns are running a performance deal while being paid a CPL rate, and they usually do not realise it until the month closes.

Setting windows per reason, not globally

Single global windows are why insurance lead return policy negotiations stall. A 72-hour blanket window is unfair to the buyer on a consumer complaint that surfaces in week three. A 30-day blanket window is unfair to the seller on a bad phone number the buyer could have detected in ten seconds.

Set the window to how long the defect could reasonably take to discover. Bad phone number, discoverable immediately, so a short window. Consumer complaint, arrives on the consumer's schedule, so a long one. Duplicate detection sits in between because dedupe runs on ingest but cross-source duplicates surface later.

A signable insurance lead return policy takes items from both columns. The negotiation is about which ones, not about whether returns exist.

Credit caps and the pricing conversation

The cap is the part of a lead return policy that keeps a bad week survivable. Without it, a bad week at a single sub-ID can produce a return request that wipes out a month, and at that point the seller's only rational move is to stop delivering. With a cap, the seller's exposure is bounded and the buyer's remedy for anything beyond it is to pause the source, which is what they actually want anyway.

Cedar Line set their cap as a monthly percentage of delivered volume, with a written rule that breaching it opened a joint review inside five business days rather than triggering a penalty. The review looked at sub-ID level data first. Nine times out of ten the problem was one publisher path, and killing it fixed the number for both parties. That is the whole point of a cap: it turns a fight into a diagnostic.

Running it day to day

A lead return policy only survives contact with a busy month if the mechanics are boring and automatic.

• Returns submitted through one channel with a structured payload, never by email thread
• Every return carries lead ID, reason code, evidence attachment, and submitter name
• Seller acknowledges inside one business day and disputes inside three
• Both sides see the same weekly report: return rate by reason code by sub-ID
• Quarterly review of the reason codes themselves, because funnels change
• Credits appear as line items on the invoice, reconciled monthly

The weekly shared report is the part of a lead return policy that people skip, and it matters more than it sounds. When both parties look at the same numbers every Monday, disputes get caught at 30 leads instead of 3,000. Cedar Line's second-generation partnerships all had it, and their return-related escalations dropped to almost nothing.

Measuring whether the policy is working

Carry Cedar Line's numbers a step further, because the cap argument only makes sense with arithmetic attached. Say the buyer takes 4,000 leads a month at an $18 CPL. That is $72,000 of monthly spend. The disputed week ran a 14 percent return rate across the whole month, so 560 leads, or $10,080 of credit requested in one go. Illustrative figures, but the shape is what matters: an unbounded return right turns a routine quality wobble into a five-figure argument between two companies that were getting along fine in October.

Now apply a 6 percent monthly cap. The buyer's automatic remedy is 240 leads, or $4,320. Everything above that does not vanish; it triggers the joint review inside five business days. In Cedar Line's case the review found one sub-ID responsible for 71 percent of the flagged volume. Killing that path cost the seller about 300 leads a month of gross volume and saved the account. The cap did not decide who was right. It bought both sides the five days they needed to find out.

Three numbers tell you whether a lead return policy is healthy, and you should publish all three monthly.

• Return rate by reason code. A rising rate on one code is a defect. A rising rate spread evenly across all six is usually a buyer process change, not a seller quality change.
• Dispute rate. The share of returns the seller contests. Under 5 percent means your evidence standard is clear. Over 20 percent means the codes are ambiguous and need a rewrite.
• Time to resolution. Median days from submission to credit or rejection. When this drifts past a week, somebody has quietly stopped doing the work, and the backlog is where relationships rot.

Watch the direction more than the level. A 4 percent return rate that has doubled in two months is a worse signal than a stable 8 percent, because the stable number is already priced into the deal and the moving one is not. If your lead return policy has been in place a quarter and you still cannot produce these three numbers on demand, you do not have a policy. You have a paragraph.

What does not work

Verbal agreements about "being reasonable." Reasonable means different things to a CFO and a media buyer.

Unlimited returns as a sales concession. Sellers offer this to win a deal and then discover the buyer's CRM auto-returns anything not reached in three dials. You have sold a performance product at a CPL price.

No cap plus a long window. The seller cannot forecast revenue and will quietly divert their best traffic elsewhere, which the buyer will experience as declining quality without understanding why.

And blanket clauses copied from a template. Your filters, your states, your dedupe logic. A generic insurance lead return policy fits nobody's actual order.