JobMoney Broker LLC · USA
FINLEADGEN
trafficsales
Lead generation for US consumer lending: from concept to production in five months.
In short
Launched a B2B2C lead-generation platform for the US market from scratch, and new ping/post auction logic raised application-to-sale conversion from 31% to 75%.
- Platform from concept to production in 5 months: landing pages, application forms, a ping/post auction, API delivery to lead buyers.
- Application-to-sale conversion 31% → 75%: reworked routing and integrated 5 lead buyers.
- Landing conversion 35% → 50% on paid traffic through systematic funnel A/B tests.
- Analytics on PostgreSQL and Metabase, plus TCPA, FCRA, CCPA and CFPB compliance down to the state level.
31%75%+142%
application-to-sale conversion
35%50%+43%
landing: view → application
5 mo
concept to production
Context
Someone in the US looks for a loan and fills in an application on a landing page, and that application becomes a lead. Within a fraction of a second the platform offers it to buyers: lenders and networks. That is ping/post: buyers first see an anonymised “ping” and name a price, then the winner receives the full data in the “post”.
Next to it sits BrokerMoney, a personal-finance Android app and web platform. The team is distributed: engineers, designers, copywriters and freelancers.
Problem
Only 31% of applications turned into a sale. Every unsold application is a paid click burned: traffic is bought up front, and money only comes in for a lead that sells.
- Routing ignored which buyer would actually accept which borrower profile.
- Declined traffic was not monetised at all.
- The landing page turned 35% of views into applications, an expensive entry into the funnel.
Hypotheses
- 01
Route by likelihood of purchase
Sending a lead where it is most likely to be bought, instead of down a fixed queue, raises the sale rate.
- 02
More, and more different, buyers
Buyers with different credit boxes will take the applications nobody takes today.
- 03
A second life for a decline
A declined borrower can be offered an adjacent product: credit repair, debt relief, a credit builder.
- 04
The form decides
Form steps and the offer on the landing page move the top of the funnel most, so they get tested first.
Solution
- Reworked the ping/post auction logic: auction, routing and postback statuses over API.
- Integrated 5 lead buyers, including LeadsMarket, RoundSky, LeadStackMedia and CashPotUSA.
- Launched paid channels, Facebook Ads and Reddit Ads, with unit economics under control: EPL, CPA, EPC.
- Designed declined-traffic monetisation so that EPL grows beyond the primary sale.
Try it
Data and compliance
I built the analytics stack on PostgreSQL and Metabase: dashboards for EPL by tier, partner efficiency, requested-to-paid and UTM/sub-ID attribution. Routing and traffic-buying decisions rest on it; without it a win is indistinguishable from noise.
Financial leads in the US are heavily regulated: TCPA, FCRA, CCPA, the FTC Act and CFPB rules. I designed and shipped state-level routing restrictions: a lead physically cannot go to a buyer who is not allowed to buy it.
AI is part of the loop: landing and app prototypes, generated specs and reports, Python and JS data scripts, all in Claude Code and Cowork.
Impact
31%75%+142%
application-to-sale conversion
35%50%+43%
landing: view → application
5 mo
concept to production
5
lead buyers integrated over API
What I take forward
- Measure first, optimise second: dashboards by tier show where money leaks better than any debate.
- Regulation is part of the product, not a legal footnote: state rules live right inside the routing.
- A decline is a segment too. A borrower who was not approved still has a financial problem to solve.
Want more? Switch the reading depth to 30 min.