PushFire

Selected engagements

What the records showed.

Two anonymized accounts, reviewed against source data the client can inspect. Each case includes what worked, what did not, and what changed as a result.

Generate · Optimize

Three times the leads. Roughly half the total marketing outlay.

National employment firm

Lead volume
3.0×

April–June 2025 vs. April–June 2026

Total monthly marketing outlay
−49%

Comparable months YoY (peak-month reference stated below)

Cost per lead
−56%

Comparable months YoY

Organic conversion rate
1.27% → 2.03% (1.6×)

12 months rolling

Editorial referring domains per month
2.22 → 17.25 (7.8×)

Monthly average, Ahrefs 44-month backlink history

Editorial = external publisher links; excludes directories, syndicated releases, owned sites, paid placements, and 538 spam-flagged domains discarded from 825 total.

From April through June 2025 to the same months in 2026, monthly marketing outlay fell and lead volume roughly tripled year over year. The late-2024 peak month for reference: $16,677 in total marketing outlay. The current monthly run rate is approximately $8,500.

What changed on one practice-area page

One practice-area page was rebuilt in October 2025. In the 12 months after the rebuild the page generated 190 leads, compared with 123 in the prior 12 months, on comparable recorded traffic. Organic conversion rate on the page moved from 1.27% to 2.03%.

That is an observation, not a controlled experiment — search demand, brand recognition, tracking coverage, and other work all move in the background. What the account records support directly: the page continued generating leads without incremental media spend during the period measured.

Verified across three separate systems

  1. The firm’s analytics for sessions and conversion.
  2. The firm’s answering service for calls that qualified as actual matters, matched back to the CRM on caller number.
  3. Ahrefs, an external count of referring domains rather than an on-site metric controlled inside the firm’s analytics account.

What the account records could not yet prove

Of the 73 tracked inbound calls in the review window, 18 (24.7%) were abandoned before anyone answered, at a median wait time of 46 seconds within that abandoned subset. A LinkedIn test spent $442 for 23 clicks — a $19.22 CPC against Google’s $6.84 per click in the same window. On whether the phone and form records could be combined into a single count of signed matters, the report concluded that a combined count was not yet supportable from the available records.

Generate

Most classified organic-search traffic was branded.

Event venue, Houston market

Unpaid search — nonbrand only
~2,400 visits → ~80 inquiries (estimate)

12 months rolling

Estimate derived from form submissions and tracked calls attributed to unpaid nonbrand search; calls without a reliable source were excluded from the estimate.

Google Ads (no brand terms bid)
4,464 visits → 111 inquiries

12 months rolling

Facebook + Instagram ads
17,301 visits → 3 inquiries

12 months rolling

Share of classified organic search that was branded
74%

12 months rolling

Google Search Console data classified as brand, nonbrand, or unclassified; omitted and anonymized queries excluded from the classification.

Recommended Facebook budget reductionrecommendation
−80%

Recommendation — redeploy to remarketing on already-visited users.

The client’s question was reasonable: if more than half of recorded inquiries were attributed to unpaid search, why pay for ads at all?

Of the queries Search Console could classify, 74% contained the venue’s name. Most classified organic traffic therefore reflected existing brand awareness rather than nonbrand discovery. Once brand searches were stripped out, the comparison changed. On a like-for-like nonbrand basis, Google Ads reached 1.86× as many visitors as unpaid search — and generated more inquiries from them (111 vs an estimated ~80). Facebook produced one inquiry per 5,767 visitors — roughly 0.017%.

Facebook produced three inquiries from 17,301 visits. The report recommended reducing the Facebook budget by 80% — roughly $472 a month — and redeploying to remarketing on already-visited users. Continue investing in Google Ads, which produced a fifth of recorded inquiries. PushFire managed the Facebook channel.

All 549 inquiries reviewed by hand

Every inquiry was read individually rather than sampled. 97% carried plausible names and complete phone numbers, with a specific request in the first message. 48% arrived ready to talk — defined as an inquiry that named a date, guest count, or budget in the first message.

A live spot check surfaced what the dashboard could not

The client’s third-party listing platform sent leads that never touched their website, so no analytics could measure them. In one live test, we submitted an inquiry through it. It was distributed to four venues simultaneously, all four preselected, with the message written by the platform rather than the couple and the guest count the platform already knew stripped out. One observation — not a survey of platform behavior — but enough to establish a question the client’s dashboard could not.

Client names are withheld pending written approval. Figures are drawn from the reports delivered to each client; rounding and explanatory labels are noted where used.

Bring us the account you cannot afford to misread.

We'll review the source accounts, identify what can and cannot be proved, and recommend where money should move.

Inquire