When we launched the Coverly embed widget, we expected brokers to drop it on their homepage and call it done. Six months of partner data later, the picture is more interesting: brokers who got the most lift from the embed did three very specific things, and brokers who did not move much did three other very specific things. This post walks through the patterns we saw, the numbers behind them, and the placement decisions that mattered the most.
What we measured
Across the partner panel, we tracked four numbers per embed instance:
- Form submit rate — what percentage of embed impressions resulted in a submitted form.
- Time-to-quote — median seconds between form submit and the producer seeing a structured quote in their inbox.
- Quote-to-bind — of the leads the embed produced, what percentage bound within 30 days.
- Attribution clarity — did the broker's CRM end up knowing which submit came from which source-event?
The fourth number is the one most broker operators underestimate. The brokers who could finally attribute which lead source produced which bound policy started optimizing their marketing spend in a way that the brokers still running spreadsheet rebuilds could not.
Pattern #1 — Homepage entry-point + product-page depth
The biggest lift came from partners who put a hero-level "Get a quote in 60 seconds" CTA on their homepage and a full-width embed on each life-product or auto-product landing page. The CTA on the homepage is short and visual — it primes the visitor. The embed on the product page is the working surface — that is where the form actually lives.
Partners who only did the homepage CTA saw high click-throughs but mediocre submit rates; visitors bounced at the click because there was nothing to do once they arrived. Partners who only did the product-page embed saw strong submit rates but low impressions; the right visitors never made it there. Both pieces together produced the lift.
Pattern #2 — Co-branded carrier panel at the top of the results page
The second pattern came from white-label partners (typically MGAs and FMOs) who put the coverly-powered carrier panel on a substitute domain path under their own brand. This is not technically embed-as-widget, but it is embed-as-pattern: the partner owns the page, Coverly owns the underlying quote engine, attribution flows back to the partner through the contact's partner_id tag.
This pattern produced the strongest attribution clarity of any configuration we measured. Every lead was stamped to a partner on submission, the partner's branded result page drove the visitor's expectations, and the producer on the back-end had a complete attribution record on the contact row by the time the lead hit their inbox.
Pattern #3 — Producer-popup with pre-quote chat
The third pattern was a smaller-volume, higher-conversion experiment from a partner who embedded a chat-style pre-qualification step before the full quote form. The pattern: visitor lands, sees a chat widget, answers three questions about what they are shopping for, and is then handed to the full embed form pre-populated with their answers.
This lifted submit rates meaningfully on traffic that was already high-intent (paid social). Most partners reported it was not worth the engineering lift on cold traffic, so this is a niche pattern — every partner who tried it on cold traffic reverted to Pattern #1 within a month.
What did not work
Three patterns we saw partners try that produced flat or negative results, and which we now flag in our partner docs:
- Embed on the contact page. Visitors on the contact page wanted to talk to a human, not fill out a new form. Pulling them into the embed cut the contact-page conversion in half and pulled the embed's submit rate down by a third. Net negative.
- Embed as a pop-up modal on every page. We tried this on ourselves and it drove a quick spike in impressions and a quick cratering of submit quality. Modals after fifteen seconds felt fine; modals on exit-intent felt punitive. We pulled it everywhere.
- Embed with six-step progressive disclosure. One partner built their own version that broke the form into six steps expecting it to feel lighter. It felt heavier. Producers complained, consumers bounced at step three. We reverted to a single-page form.
The placement rule of thumb
If you are evaluating where to put the embed, the rule of thumb that produced the most consistent lift across our partner panel was:
The embed belongs on the page where a visitor already knows what product they want. Not the homepage. Not the contact page. The life-product landing page, the auto-product landing page, the rate-quote page. The visitor who landed there is the highest-intent visitor on the site; the embed's job is to give that visitor a path to a structured quote in under ninety seconds.
Everything else — chat widgets, modal pop-ups, six-step forms — is decoration on top of that core job. Decorations can lift by two or three percentage points. The base pattern is doing ninety percent of the work.
What we would build differently if we started over
Honestly: not much. The embed widget is six files and one React component. The connector layer reads the same way. The qualification agent is a four-step pipeline. The patterns we see working at scale are unglamorous — hero-level CTA, full-width embed on product pages, results on a co-branded panel, attribution pinned to partner_id on the row.
The work for the next six months is to make the unglamorous version faster and easier to deploy, not to add another shiny surface on top of it.