More accepted leads from the same click volume
Client: SolarReviews (US residential solar lead generation / email traffic program).
Channel: Email publisher traffic to accepted homeowner leads.
From the same 10,000 delivered clicks, accepted enquiries rose from 350 to 500. Accepted share of submissions rose from 53.8% to 80.6%. Campaign revenue rose from $7,000 to $10,000; contribution after publisher cost rose from $2,100 to $3,000.
Client & brief
SolarReviews buys accepted residential-solar enquiries through an email publisher arrangement. Publishers retained their subscriber lists and sent to their own audiences. The commercial arrangement covered traffic and accepted enquiries, not the sale of a subscriber database.
An accepted enquiry was defined in writing as: homeowner, supported state, valid contact details and no duplicate within the agreed period.
SolarReviews payment: $20 per accepted lead. Publisher payment: $14 per accepted lead. Remuneration in this engagement was based on accepted enquiries. A cost-per-click purchasing arrangement would have different economics.
Situation before engagement
An initial email campaign delivered 10,000 clicks to SolarReviews’ enquiry page. It produced 650 form submissions, but only 350 were accepted under the campaign’s qualification rules.
Rejection breakdown: 120 enquiries outside the supported geography, 80 from people who did not own the property, 60 duplicates and 40 with invalid contact details. Each rejected lead was counted once using its primary rejection reason.
At first glance, the campaign converted 6.5% of clicks into forms. The accepted-lead rate was only 3.5%, so advertiser feedback mattered more than the raw submission count.
What we changed
Each publisher received a separate tracking identifier so the source of rejected enquiries could be reviewed.
The email and destination page made the homeowner requirement clearer. Publisher selection focused on the agreed geographic audience. Form checks flagged unsupported locations and obvious duplicates. Publisher-level feedback identified traffic that needed to be paused and investigated.
Results
Two equal-volume campaign cohorts, each reviewed over the same 30-day acceptance window.
| Measure | Before | After |
|---|---|---|
| Delivered clicks | 10,000 | 10,000 |
| Submitted enquiries | 650 | 620 |
| Accepted enquiries | 350 | 500 |
| Rejected enquiries | 300 | 120 |
| Accepted share of submissions | 53.8% | 80.6% |
| Click-to-accepted-lead rate | 3.5% | 5.0% |
| Advertiser payment per accepted lead | $20 | $20 |
| Publisher payment per accepted lead | $14 | $14 |
| Campaign revenue | $7,000 | $10,000 |
| Publisher cost | $4,900 | $7,000 |
| Revenue less publisher cost | $2,100 | $3,000 |
What the numbers mean
Accepted leads rose 42.9% while submitted enquiries fell slightly. The improvement was in acceptance quality, not a larger top-line form count.
Revenue rose by $3,000, but publisher costs also rose because the publisher is paid for each accepted lead. Contribution before other costs rose by $900. The margin on revenue remained 30%; this was not a margin-expansion story.
The $3,000 after publisher costs is before tracking, staff, administration and other overhead. It is not net profit, personal income or solar installation revenue. Installer quotes, signed installations and cash collections sit downstream of this engagement.
Operating arrangement
Three parties: the advertiser buying accepted solar enquiries; publishers sending to their own email audiences; and Meridian coordinating the commercial arrangement and reviewing source-level feedback.
The advertiser specified supported geography and acceptance rules before the test. Publishers agreed how accepted leads would be counted, the review window and the payment basis. Campaign links carried a source identifier so results could be reconciled without transferring a publisher’s subscriber list.
Where the quality problem appeared
Three sources, labelled Publisher A, B and C for anonymity:
| Baseline source | Clicks | Submitted leads | Accepted leads | Accepted share of submissions |
|---|---|---|---|---|
| Publisher A | 4,000 | 270 | 180 | 66.7% |
| Publisher B | 3,500 | 220 | 130 | 59.1% |
| Publisher C | 2,500 | 160 | 40 | 25.0% |
| Total | 10,000 | 650 | 350 | 53.8% |
Publisher C had the weakest accepted-lead result. That was a reason to investigate, not proof of fraud. Possible explanations included audience mismatch, unclear eligibility wording or a problem in the journey between the email and form.
Investigation and changes
First review: Reconcile submitted leads against the advertiser’s accepted and rejected records. Check that a duplicate lead has not been counted twice and that rejection reasons are applied consistently across sources. Resolve tracking discrepancies before making a commercial claim about quality.
Source review: Compare geography and rejection categories by publisher. Ask each publisher to explain the audience available for the campaign and the promotion used. Pause further volume from Publisher C while reviewing the mismatch, rather than continuing to buy because click delivery is on schedule.
Message and form review: Make the homeowner requirement and geographic availability clear before the visitor submits the form. Check that campaign links retain their source identifiers. Keep form language consistent with the advertiser’s actual offer and contact process, without adding unsupported savings or eligibility claims.
Controlled restart: Start Publisher A and B with a 1,000-click test allocation each. Review acceptance feedback before committing the remaining volume. Final allocation for the comparison cohort: 6,000 clicks to A and 4,000 to B; C remained paused.
Commercial reconciliation: Review provisional counts weekly, then close the cohort using the agreed 30-day acceptance window. Accepted-lead totals, rejection reasons and any unresolved disputes reconcile before payout is finalised. This cohort closed with no unresolved adjustments.
Final volume distribution
| Revised source | Clicks | Submitted leads | Accepted leads | Accepted share of submissions |
|---|---|---|---|---|
| Publisher A | 6,000 | 380 | 320 | 84.2% |
| Publisher B | 4,000 | 240 | 180 | 75.0% |
| Publisher C, paused | 0 | 0 | 0 | Not applicable |
| Total | 10,000 | 620 | 500 | 80.6% |
The result combines a change in source mix with clearer eligibility messaging and form checks. It does not isolate the contribution of each intervention. The improvement should not be attributed solely to pausing Publisher C.
Rejection reasons after the changes
| Primary rejection reason | Before | After |
|---|---|---|
| Outside supported geography | 120 | 35 |
| Not the property owner | 80 | 30 |
| Duplicate enquiry | 60 | 35 |
| Invalid contact details | 40 | 20 |
| Total rejected | 300 | 120 |
The after cohort still had rejected enquiries. Screening and clearer messaging reduced avoidable mismatches, but did not make every visitor eligible or every contact detail usable.
Revenue, margin and cash
At 500 accepted enquiries, campaign revenue was $10,000 and publisher cost $7,000. The remaining $3,000 is contribution before other costs. Tracking services, operations time, payment costs and overhead still need to be deducted to reach operating profit.
Payment terms in this engagement: advertiser pays 15 days after final acceptance; publishers are paid seven days after final acceptance. That creates an eight-day funding gap. For the after cohort, the $7,000 publisher payment was covered before the advertiser’s $10,000 was received. Those timings are separate from the contribution figure above.
An accepted enquiry is not a booked survey or a signed solar installation. Without downstream feedback, reporting supports statements about lead acceptance and contribution after publisher cost, not about solar sales or homeowner outcomes.
Source-level reporting and follow-up
Operating reports included campaign reference, publisher identifier, reporting period, delivered clicks, submitted leads, accepted leads and rejection reasons. Totals reconciled with the advertiser’s final acceptance file. Personal contact details were kept out of performance summaries shared more widely.
Publisher C generated 160 submissions from 2,500 clicks but only 40 were accepted. The follow-up requested a review of the audience and eligibility message before more volume was allocated. The publisher was not labelled fraudulent on acceptance rate alone.
Where a publisher disputed a rejection, the advertiser referred to the agreed reason and supporting record. A request that met the written criteria was not rejected merely because the installer did not close a sale, unless the commercial agreement explicitly used that later event as its acceptance condition.
Tracking issue we ruled out
An email link that redirects through a landing-page tool can drop its publisher identifier and make otherwise valid leads appear unattributed. Before pausing a source, the team followed a test link through the full journey and confirmed that the identifier reached the lead record. Test records were labelled and excluded from billable lead totals.
Next phase
The final 10,000-click cohort does not establish that a publisher can deliver the same acceptance rate at five times the volume. Larger allocations may reach a different part of an audience, create more overlap or change the mix of visitors.
Next phase: another limited order, retain source identifiers and continue reviewing advertiser feedback. Decisions consider accepted-lead consistency and downstream contactability, not only the $3,000 contribution shown here. No further sales or scale uplift is assumed in these figures.
Interpretation notes
Keep source-level click records, lead IDs, submission timestamps, acceptance decisions and primary rejection reasons. Retain agreed commercial terms and reconcile invoiced leads against the final accepted count. Protect personal data; showcase summaries only need aggregated results.
Two sequential cohorts may differ in demand and audience composition. Consistency over further orders, complaint levels, contactability and advertiser downstream feedback matter before treating the after figures as repeatable performance.