Case study 01 · Flagship

Four brands, one search strategy.

Senior Social Media Specialist for a New England behavioral-health network — Meta Addiction Treatment, Pioneer Behavioral Health, Luna Recovery for Women and Pillars Health Group. My own brands competed with each other in the same towns, in Google's most gated ad category. The job wasn't ranking — it was deciding which brand was allowed to win which query.

Apr 2022 – Sep 2024 · 29 months Healthcare · addiction treatment Local SEO · Paid · Social · Referral
Verified
4
consumer brands
Confirmed from the brands' live sites, Aug 2026.
Verified
9
sites, 2 states
Massachusetts + Maine.
Attested
+40%
patient inquiries
Internal reporting across the network.
Attested
−25%
cost per click
In a LegitScript-gated category.
Attested
+30%
partner engagement
Clinician referral network.

The network

Internally it was one marketing operation — "Behavioral Health Services" — publicly it was four consumer brands with separate sites, separate Google Business Profiles and deliberately different audiences. All four still share the site architecture and program naming I worked with, which you can verify today.

MAINE MASSACHUSETTS Longmeadow Oxford Marlborough Concord S. Portland MERRIMACK VALLEY Detail — four sites inside 20 minutes Haverhill — Meta (women) N. Andover — Luna (women) Tewksbury — Pillars N. Reading — Meta Two women-only programmes, 15 minutes apart, same SERP.

Schematic, not to scale. Locations verified from the four brand websites, August 2026. The brands referred patients to each other — pioneerbh.com still routes women's and adolescent enquiries to North Andover and Concord.

The problem nobody practices for

Anyone can do local SEO for six branches of one brand. This network was segmented by who, not by where: Luna (women only) and Meta–Haverhill (women only) sit fifteen minutes apart serving the same demographic. Without coordination, the network's own brands cannibalize each other's rankings and bid against each other in paid.

One facility, many front doors

Luna has a single building in North Andover — but ranks across a metro. I built geo-targeted landing pages extending each facility's reach into surrounding towns. They're still live: open lunarecoverycenter.com/boston-ma/ — a Boston page whose every enquiry routes to the North Andover facility.

PHYSICAL FACILITY Luna Recovery 1270 Turnpike St North Andover, MA /north-andover-ma/the real address — canonical /boston-ma/verified live · routes to N. Andover /andover-ma/verified live /lawrence-ma/same pattern, other brands /haverhill-ma/same pattern, other brands RESULT One catchment, many local SERPs without a second lease

Solid outlines = verified live, August 2026 — two years after I left. Dashed = the same pattern deployed across the network's other brands.

Why −25% CPC here is not −25% anywhere else

Addiction treatment is the most restricted category in paid search. Google refuses these ads outright unless the advertiser holds LegitScript certification — and because a patient's lifetime value is enormous, the category's CPCs are among the highest that exist. The cut was achieved inside both of those gates.

Any advertiser wants to bid on treatment keywords GATE 1 LegitScript certification is mandatory before Google will serve a single ad GATE 2 Category cost among the highest CPCs in all of paid search — patient LTV is enormous −25% cost per click inside both gates

Gates verified: Pillars Health Group carries LegitScript verification alongside Joint Commission and Massachusetts DPH accreditation. The −25% is my own measurement from internal campaign reporting.

How rank was tracked — the method

Local rank isn't one number: it changes with the searcher's position. I ran grid-rank tracking in BrightLocal around every site — dozens of overlapping grids across nine locations — to see where each brand held the local pack, where it fell to page one, and where two of our own brands were fighting each other.

Ranking varies by the searcher's location — a single "rank" is meaningless 14 8 6 17 9 3 2 7 6 1 1 4 21 10 7 15 Grid of simulated search points around one clinic What the grid shows Each point is a simulated search from that spot. Rank collapses with distance from the clinic — so "we rank #1" is true and useless. Why it mattered here Nine sites × four brands = dozens of overlapping grids. Where two of our own brands both ranked, we were bidding against ourselves. Geo landing pages extend green coverage outward without a second lease. Illustrative diagram of method — not client data. Reading the pins Top 3 — in the local pack 4–10 — page one, no pack 11+ — effectively invisible

The pin numbers are illustrative and labeled as such on the graphic itself. They explain the method; they are not client data. A healthcare employer's exports aren't mine to publish — and a faked screenshot would be worse.

What I did

BrightLocalgrid rank tracking, 9 sites
Google AdsLegitScript-gated category
Meta Business Suitepaid social, 4 brands
Google Business Profile9 listings
GA4inquiry attribution
IG · TikTok · YouTube4 brand voices

VerifiedNetwork structure, locations, geo pages, LegitScript status — from the brands' public websites, checked August 2026; links above.

Attested+40% inquiries · −25% CPC · +30% partner engagement — my measurements from internal reporting (calls + form fills; campaign CPC vs prior period; referral activity). Tool access ended with my tenure, so no exports are shown — the flags exist so you know exactly what rests on my word. Nothing patient-level was ever part of marketing reporting.