Good optimization does not flatten differences. It respects them.
Michael McBride
Founder & CEO, Atelier Commerce
BUILT WITH
Good optimization does not flatten differences. It respects them.
Michael McBride
Founder & CEO, Atelier Commerce
BUILT WITH
The same truth, told to two audiences, requires two different sentences. Our job is to know which sentence belongs to whom, and to say nothing we cannot stand behind to either one.
The same truth, told to two audiences, requires two different sentences. Our job is to know which sentence belongs to whom, and to say nothing we cannot stand behind to either one.
The Challenge
The dual-storefront model created problems that a single-store program never has to solve.
The first was language. A meta title that converts a consumer actively repels a physician. "Best Hair Growth Serum" earns the click from someone comparing options on their phone. To a dermatologist evaluating what to stock and recommend, that same phrasing reads as marketing, not medicine. The professional audience responds to clinical framing: peptide complex, in-office protocol, science-backed, measurable results. Every title tag, every meta description, and every H1 had to be written twice, in two registers, for two readers who would never see the other version. The second was duplication. Because both stores sell overlapping products built from the same source content, the catalogs were riddled with duplicate title tags, duplicate meta descriptions, and duplicate H1s across product, collection, and blog pages. Search engines cannot rank what they cannot tell apart. On the consumer store, duplicate metadata blurred the line between a blog article and a collection page. On the professional store, duplicate H1s and titles ran across multiple rounds of products, diluting the very pages meant to signal clinical authority. Left alone, the two stores would have competed with themselves and, in places, with each other. The third was measurement. There was no clean way to read either store's organic performance. Tracking was incomplete, Search Console did not exist for either property, and the data that did flow in could not be trusted to separate consumer behavior from professional behavior. You cannot optimize an audience you cannot see. The fourth was the new front of discovery. Consumers increasingly start with an AI assistant rather than a search box. A storefront that cannot be read cleanly by an agent is invisible in that channel, no matter how well it ranks in traditional results. The consumer store needed to be legible to machines as well as people.
The Challenge
The dual-storefront model created problems that a single-store program never has to solve.
The first was language. A meta title that converts a consumer actively repels a physician. "Best Hair Growth Serum" earns the click from someone comparing options on their phone. To a dermatologist evaluating what to stock and recommend, that same phrasing reads as marketing, not medicine. The professional audience responds to clinical framing: peptide complex, in-office protocol, science-backed, measurable results. Every title tag, every meta description, and every H1 had to be written twice, in two registers, for two readers who would never see the other version. The second was duplication. Because both stores sell overlapping products built from the same source content, the catalogs were riddled with duplicate title tags, duplicate meta descriptions, and duplicate H1s across product, collection, and blog pages. Search engines cannot rank what they cannot tell apart. On the consumer store, duplicate metadata blurred the line between a blog article and a collection page. On the professional store, duplicate H1s and titles ran across multiple rounds of products, diluting the very pages meant to signal clinical authority. Left alone, the two stores would have competed with themselves and, in places, with each other. The third was measurement. There was no clean way to read either store's organic performance. Tracking was incomplete, Search Console did not exist for either property, and the data that did flow in could not be trusted to separate consumer behavior from professional behavior. You cannot optimize an audience you cannot see. The fourth was the new front of discovery. Consumers increasingly start with an AI assistant rather than a search box. A storefront that cannot be read cleanly by an agent is invisible in that channel, no matter how well it ranks in traditional results. The consumer store needed to be legible to machines as well as people.
The System
We ran two SEO campaigns in parallel, one per store, and refused to copy answers between them.
On the consumer store, we resolved duplicate meta descriptions and title tags across successive rounds, introducing page-type differentiators so a "Blog" page and a "Collection" page could no longer be confused in search. We standardized a reusable meta protocol (product title, keyword, and brand name) applied consistently across products, collections, content, and blog pages, so every page now states plainly what it is and who it is for. We also published an llms.txt file so AI agents and shopping assistants can read the store directly, find products, and understand how to transact, turning the storefront into something answer engines can quote rather than skip. In the professional store, the same problems demanded different answers: we rewrote duplicate H1s, title tags, and meta descriptions in medically credentialed language built for physicians, dermatologists, and clinic providers. Where the consumer store says serum, the professional store says protocol; where one promises visible results, the other documents clinical outcomes, the same product deliberately described in two vocabularies so each store earns the right audience. Across both stores, we cleaned up the foundation. We remediated 404 errors surfaced in Search Console so authority stopped leaking through broken paths, and built Google Search Console from scratch for both properties, giving each store its own honest view of impressions, clicks, and queries. We deployed a Littledata script to feed clean purchase and behavior data into GA4, and added Microsoft Clarity for heatmapping and session recording, so we could watch how each audience actually moved rather than guess. Then we turned to conversion, building systems that work without supervision. On the professional store, we optimized the login redirect so a tagged B2B customer lands ready to reorder, restructured the Laser Cap product page to cross-sell the KeraFactor serum, and stood up a Recharge subscription portal for reorders on their own cadence. On the consumer store, we put Shopify Flow to work on fraud detection, screening risky orders automatically. We built dedicated blog and press landing pages to capture and convert the 858M-plus impressions of 2025, migrated SMS from Attentive to Klaviyo to unify messaging with email, and built HubSpot lead-nurture workflows tuned to the longer B2B buying cycle. A migration from Stamped to Reviews.io and Influence.io is underway, rebuilding loyalty and rewards so social proof and repeat purchase compound rather than scatter.
The System
We ran two SEO campaigns in parallel, one per store, and refused to copy answers between them.
On the consumer store, we resolved duplicate meta descriptions and title tags across successive rounds, introducing page-type differentiators so a "Blog" page and a "Collection" page could no longer be confused in search. We standardized a reusable meta protocol (product title, keyword, and brand name) applied consistently across products, collections, content, and blog pages, so every page now states plainly what it is and who it is for. We also published an llms.txt file so AI agents and shopping assistants can read the store directly, find products, and understand how to transact, turning the storefront into something answer engines can quote rather than skip. In the professional store, the same problems demanded different answers: we rewrote duplicate H1s, title tags, and meta descriptions in medically credentialed language built for physicians, dermatologists, and clinic providers. Where the consumer store says serum, the professional store says protocol; where one promises visible results, the other documents clinical outcomes, the same product deliberately described in two vocabularies so each store earns the right audience. Across both stores, we cleaned up the foundation. We remediated 404 errors surfaced in Search Console so authority stopped leaking through broken paths, and built Google Search Console from scratch for both properties, giving each store its own honest view of impressions, clicks, and queries. We deployed a Littledata script to feed clean purchase and behavior data into GA4, and added Microsoft Clarity for heatmapping and session recording, so we could watch how each audience actually moved rather than guess. Then we turned to conversion, building systems that work without supervision. On the professional store, we optimized the login redirect so a tagged B2B customer lands ready to reorder, restructured the Laser Cap product page to cross-sell the KeraFactor serum, and stood up a Recharge subscription portal for reorders on their own cadence. On the consumer store, we put Shopify Flow to work on fraud detection, screening risky orders automatically. We built dedicated blog and press landing pages to capture and convert the 858M-plus impressions of 2025, migrated SMS from Attentive to Klaviyo to unify messaging with email, and built HubSpot lead-nurture workflows tuned to the longer B2B buying cycle. A migration from Stamped to Reviews.io and Influence.io is underway, rebuilding loyalty and rewards so social proof and repeat purchase compound rather than scatter.
Success Stories
Scarlett Gasque: The Open Was Never the Problem
A Klaviyo program that looked healthy on every surface metric was quietly leaving revenue on the table. We found the leak in a single number and fixed it in 90 days, without a rebuild. Sessions fell 32%. Revenue rose 78%. The brand didn't get more visitors. It got better ones, and converted them.
Scarlett Gasque: The Open Was Never the Problem
A Klaviyo program that looked healthy on every surface metric was quietly leaving revenue on the table. We found the leak in a single number and fixed it in 90 days, without a rebuild. Sessions fell 32%. Revenue rose 78%. The brand didn't get more visitors. It got better ones, and converted them.
PXG × Nosto: 104 Days of Compounding Efficiency
PXG activated a full Nosto personalization stack across every major page type of a mature Shopify storefront. In the 104 days that followed, conversion rate rose nearly 25% and revenue per session nearly 28%, on 20% less seasonal traffic than the window before it.
PXG × Nosto: 104 Days of Compounding Efficiency
PXG activated a full Nosto personalization stack across every major page type of a mature Shopify storefront. In the 104 days that followed, conversion rate rose nearly 25% and revenue per session nearly 28%, on 20% less seasonal traffic than the window before it.
Success Stories
Scarlett Gasque: The Open Was Never the Problem
A Klaviyo program that looked healthy on every surface metric was quietly leaving revenue on the table. We found the leak in a single number and fixed it in 90 days, without a rebuild. Sessions fell 32%. Revenue rose 78%. The brand didn't get more visitors. It got better ones, and converted them.
PXG × Nosto: 104 Days of Compounding Efficiency
PXG activated a full Nosto personalization stack across every major page type of a mature Shopify storefront. In the 104 days that followed, conversion rate rose nearly 25% and revenue per session nearly 28%, on 20% less seasonal traffic than the window before it.
