Case studies & reviews

A quote request on 96 pages, up from one. Name, phone and email on every one, and the full version on the contact page takes a photo of the room and routes the message by what the customer picked.

Thirty separate stylesheets on the old homepage, and 258 KB of code before any content. The rebuilt page uses one stylesheet and 33 KB. Same words, a fraction of the weight.

Twenty years of their blog moved across intact and now sits in an editor their own staff can use, so a post can be written without touching the code the site is built from.

The only two addresses on the site that did not exist before: a careers page, published with the job markup that makes a role eligible for Google Jobs, and a privacy policy, which they did not have.

Every page ends in a quote request. The date picker knows the shop is closed weekends, and a customer can attach a photo of the fireplace they want replaced.

A page each for Jøtul, Real Fyre, White Mountain Hearth, Osburn and Buck Stove, so a search for the brand can land on the dealer who stocks it.

Every page carries a plain-English FAQ backed by FAQPage markup, so the answers are eligible to show up in Google rather than sitting buried in a paragraph.

A plain-text summary for AI search, plus explicit permission for the ChatGPT, Claude and Perplexity crawlers, so the shop can be named when someone asks an AI where to buy a fireplace in Rome.

A quote request on 23 of the 27 pages. Name, phone and email required, a date picker for planning the visit, and a drop zone for photos of the space, because sizing starts with seeing it.

A page each for Napoleon, White Mountain Hearth, Majestic, Summerset, Wildfire, MasterCool and Hessaire, so a search for the brand can land on the dealer who stocks it.

Plain-English answers across eleven pages, each backed by FAQPage markup, so they are eligible to show up in Google instead of sitting buried in a paragraph.

Their own fire, filmed in their own store, converted and compressed to run silently on a loop. Almost every fireplace image online is a computer rendering. Theirs actually burns.

Of tracked calls connected to the store: 969 unique callers, and 54% of calls ran 2+ minutes. These are conversations, not wrong numbers.

Of search leads came from Microsoft Bing, a second search engine most agencies never even set up.

Of the year's 299,223 ad impressions came from social running on under 4% of the budget, cheap brand presence between purchases, while search does the converting.

The single busiest call hour of the week. Reporting this granular tells you when the phone needs a person on it.

No ramp-up period. The campaign launched September 30 and produced 79 leads in October: proof that a well-built launch can meet the season instead of chasing it.

Search CTR: 1,638 clicks on 16,550 search impressions. Tight targeting means the ad only shows when it's relevant and gets clicked when it shows.

Of leads were phone calls: 135 unique callers, every tracked call connected, busiest window Monday mornings. Exactly what a consultative showroom wants.

Homeowners reached by the mobile retargeting layer, staying in front of considered-purchase shoppers between showroom visits, on a small slice of the budget.

Cost per lead by the season's peak, down from $117 in the first full month. This is what the XMO learning curve looks like when it's given a head start before the season.

Of 712 tracked calls connected: 473 unique callers averaging 3 minutes 11 seconds, and 44% of calls ran 2+ minutes.

Of ad clicks came from mobile, where a shopper standing in a competitor's showroom does their comparison searching.

Leads arrived as phone calls (712), qualified website actions (90), and emails (14), every one tracked back to the campaign, so nothing gets credited on faith.