SGA Website Ecosystem — Platform Direction

1The Situation

SGA received two competing proposals for its website portfolio, and neither was built with visibility into work already on SGA's own roadmap.

Plan	Owner	Core Idea	Status
WP Multisite Consolidation	OM Marketing (external vendor)	One shared WordPress theme, multisite codebase, hosting savings	Proposal in hand
Foundation Engagement	Nic Clark (independent advisor)	Build a data repository + portfolio-wide SEO/analytics quick wins before deciding platform	Proposal in hand ($15K–$55K tiers)
WS1-A4 Storefront Generation	Internal (SGA roadmap)	JAMstack site-generation engine — replaces or integrates with existing sites	Already specified, not yet started

SGA is not choosing between the OM and Nic proposals as written. The internal JAMstack direction is confirmed as the target architecture — both external proposals are being evaluated only as potential inputs, not as the platform decision itself.

2The Recommendation

Build the JAMstack site-generation engine in-house, owned and operated by Don Clark and Wyatt. SGA keeps 100% ownership of the codebase, hosting, and CMS layer — no vendor owns the platform.

Why this doesn't require a vendor to build it

A website platform needs five things. SGA already has four of them running or decided:

Need	Status
Brand / design-token system	✅ Already exists — per-practice color tokens, typography, logo, tone (260 practices)
Practice / location identity source	✅ Already live — SGA Master Crosswalk API, 295 practices
Media / asset pipeline	✅ Already built and running — Asset Registry, S3 + CloudFront + image processing
Content authoring / approval	✅ Already built — Content Engine v1.0
Hosting infrastructure	✅ Already the settled choice — AWS S3 + CloudFront
Site-generation / template engine	THE ONE GAP Not yet built — the actual scope of this initiative

If an outside vendor built a WordPress Multisite instead, it would need its own brand config, its own media library, and its own hosting — a second, disconnected system duplicating what SGA already runs.

3What JAMstack Is

JAMstack = JavaScript, APIs, Markup. Pages are pre-built as static files once, then served instantly from a CDN — instead of a server assembling each page live on every visit, the way WordPress works today.

WordPress (today)

Cooks every order from scratch

Live database query on every single visit
Per-site PHP hosting + database
Constant plugin security patching
Larger attack surface

JAMstack (target)

Bakes a batch ahead of time

Static files served instantly from a CDN
No live database reachable publicly
Cheaper to host at 251-site scale
One template system, per-practice branding via data

4Portfolio Snapshot

251 sites: 143 SGA, 93 Gen4, 15 Modis brand.

WordPress + Divi

84

WordPress + Other

65

WordPress + Elementor

44

Unknown platform

31

Dobson Digital

8

Officite

7

Wix / Squarespace / Duda / other

9

PBHS / ProSites

3

5Migration Priority

TIER 0

Prerequisites. Resolve 31 unknown-platform sites, verify domain ownership (GoDaddy — confirmed for Gen4, in progress for SGA/Modis), validate medium-confidence detections.

TIER 1

27 sites — true vendor lock-in. Dobson Digital, Officite, PBHS, ProSites, Wix, Squarespace, Duda. No portable codebase today. Shared-domain clusters and Specialty-category practices prioritized first.

TIER 2

128 sites — WordPress, builder-incompatible. 84 Divi + 44 Elementor. Portable CMS, but the two builders don't share content formats — extraction required either way.

TIER 3

65 sites — closest to in-house-able. Plain WordPress, no proprietary builder — best near-term candidates for the new component system.

6Vendor Scope — What's Remaining

Nic Clark

Narrowed to ~$15K (from $55K Option C)

KEEP Practice archetype classification
KEEP Migration-risk sequencing
KEEP Provider/location relationship modeling (Coastal Jaw floating-provider vs. fixed-location)
DROP SEO / analytics / AI visibility — deferred to a separate initiative
DROP Domain/DNS audit — largely resolved internally

OM Marketing

Content-migration labor only

KEEP Extract copy/images/structure from 251 live sites
KEEP Priority: Elementor → Divi → plain WP
KEEP Manual capture for the 27 vendor-locked sites
DROP Theme/template design
DROP Hosting / multisite architecture

7Cost & Tools Snapshot

Bucket	Estimate	Notes
Incremental AWS infra	<$150/mo	Rounding error against platform's existing ~$10,900/mo full-scale AWS bill
Don + Wyatt + IT labor	Existing headcount	No new spend to operate; cost to build depends on an open role question
Nic Clark (narrowed)	~$15K one-time	His published Option A tier
OM Marketing	TBD	Quote pending against corrected content-migration scope

Stack direction: Node.js/TypeScript throughout (matches existing platform). Static site framework — Next.js or Astro, not yet decided. Hosting on AWS S3 + CloudFront (already the settled platform-wide choice).

8Open Questions

1Build vs. operate — Don/Wyatt have confirmed capacity to manage the tool. Is the site-generation engine itself built by them, an internal engineer, or a contractor?
2SGA/Modis domain verification — Wyatt to confirm the remaining 165 sites' registrar/DNS ownership (Gen4's 93 already confirmed in GoDaddy).
3Vendor contract end-dates — needed for the 7 proprietary-platform vendors (Dobson, Officite, PBHS, ProSites, Wix, Squarespace, Duda) to finalize migration sequencing.
4Framework decision — Next.js vs. Astro for the static site generator, pending Don/Wyatt + IT input.
5OM Marketing response — reframe their bid toward content-migration labor, share the corrected scope.