Platform Roadmap & Planning Budget
Modernize the Midwest Production Supply Platform
Fix the pricing data, rebuild on a fast Gutenberg foundation, layer in the new shopping experience, then open up AI — each phase landing on a platform that can carry it.
The store works, but it has drifted. Pricing and drop-shipper costs are hard to manage, the Divi page builder loads 84 plugins and weighs on every page, and years of custom code have accumulated. This roadmap fixes the pricing foundation, rebuilds the site for speed on Gutenberg, then adds the new shopping experience and an AI layer — sequenced so nothing is built on ground that can't hold it.
These are planning ranges, not a quote. Each phase is scoped lean, with room to expand once the foundation is in place. Software subscriptions are billed at cost and shown separately from MRA work.
Where the effort goes
Phases at their planning range. The AI layer is a menu of independent options, priced separately.
Phase detail
Pricing first; it is the prerequisite. The theme rebuild depends only on the pricing structure. The new shopping experience and AI build on the finished, fast theme.
Pricing & Cost Foundation
One-time · Before all other work
- Two-way Google Sheet, with filtering. A sheet MPS controls that both pushes and pulls, organized so purchasing can filter by brand — the way drop-shipper cost updates actually arrive.
- Push — prices & cost of goods. Update regular price and cost of goods in bulk from the sheet, writing to WooCommerce's native fields (
_cogs_value). - Pull — full product sync. Pull every product into the sheet, adding any that aren't there. Columns: ID, SKU, Publish Status, Name, Brand, Price, Cost of Goods.
- Filter & update by brand. Drop-shipper cost sheets come in by brand, so the sheet filters on the brand field for fast bulk updates.
QuickBooks → Data File (Working Session)
Optional · a working session that ends with data flowing
- A 2-hour working session that defines and connects. Together we define exactly what QuickBooks data MPS wants, show the connection possibilities through Coupler.io, and make the connection — so by the end of the session QuickBooks data is flowing into the full MPS Data File that the dashboard reads from.
- Plan the next step live. With the data in hand, we scope the follow-on hours needed if MPS wants the Looker dashboard updated to surface it — and talk through any additional needs for that financial data.
- Fits here because it touches cost and financial data; a full, ongoing QuickBooks integration is its own future scope (see boundaries).
Theme & Front End
Depends on the pricing structure · The core of the project
- Speed and usability are the goal. Remove Divi, replace it with a lightweight Gutenberg theme. Every decision serves page speed and ease of use — getting MPS onto a solid, fast foundation.
- A soft transition on staging. Load a full copy of the site to a staging environment and build there, so the customer never feels they've landed on the wrong site — the look and trust carry over while the platform underneath changes.
- Cut the plugin load. The site runs 84 plugins plus 9 must-use plugins today. We audit and reduce — for subscription savings, but primarily for speed, fewer conflicts, and less to break.
- Evaluate the custom code. A fair amount of custom code exists; we review what genuinely needs to move to the new site and rebuild it cleanly rather than carrying dead weight.
- Preserve SEO. Keep the brand and URL structure intact with a redirect plan — organic drives roughly a quarter of revenue.
UX & Communications
Builds on the fast foundation · Where the new experience begins
- Category & shopping-experience optimization. Study GoToPAC and rework the categories so a technical buyer finds the right product fast — shop-by-attribute browsing, cleaner drill-downs, faceted filtering across 2,000 products.
- Category browsing + quote request. Bring the quote-driven path into the browsing experience, so the highest-value cleanroom and configuration requests are one step from the catalog.
- Email marketing setup. Stand up Mailchimp or Klaviyo for the ~9,000 current subscribers, with WooCommerce sync, segmentation, and reporting.
- Automations planning session. A working meeting to design the first automated flows — abandoned cart and dormant-customer reactivation — before building them.
A Menu of AI Possibilities
Built on the WP Engine AI Toolkit · options are independent · requires the Gutenberg theme from 1B
These are independent options, not a fixed build — MPS can pick any, in any order. Most run on WP Engine's AI Toolkit, which keeps AI off the site's PHP workers and grounded in the real catalog. Software is billed at cost; MRA setup is the work to configure and fully utilize it.
Option A · AI Smart Search + Recommendations
Fastest winUpgrade site search from keyword-matching to AI that understands intent and spec — a buyer searching "ISO 5 wipe" finds the right SKU. Related-product suggestions are included at no extra cost. Runs on a dedicated search server, so it never slows the site.
Option B · Chat — evaluate & pilot
Test liveStand up and test a chat solution together. Tidio (a customer-service platform with AI, if they want agents + AI + a support inbox), Chatbase (streamlined, chat-only), or WP Engine's native chatbot framework grounded in the catalog. Pilot one rather than commit blind.
Option C · AI-assisted SEO & content
Runs in parallelFold AI into the content work — consistent, spec-rich descriptions and metadata across the catalog, structured so MPS surfaces in AI-assistant answers. Feeds the AI-referral channel that already shows the highest revenue-per-session.
Option D · Advanced agentic integration
Full potentialThis is the one that shows what the platform can really do — where the assistant stops answering and starts acting. Not a chatbot; a capability layer we can build almost anything on.
- Lead forms answered inside chat. The long cleanroom quote form becomes a short conversation — the assistant asks only what it needs, pre-fills from the signed-in profile, and hands the team a clean, structured quote.
- "Add my last order." A logged-in customer can reorder a previous order, add items by description, or have the assistant build the whole cart — it takes the action, not just points the way.
- Reordering & cart-loading for repeat consumables buyers — the 56% of revenue that comes back — made effortless.
- Session-scoped & safe. The assistant sees the customer's profile and history only for that conversation, with no standing data exposure.
Why it matters: once this layer exists, new capabilities are configuration, not new builds. It's the tooling that lets MPS do anything it can dream up.
Please note: phases don't need to sit at the same level, and the AI layer is à la carte. A fast, solid foundation (1A + 1B) with the new category experience (2) is the recommended core; AI options layer on as MPS is ready.
Future suggestions
Beyond the platform build, the work that keeps qualified demand growing. Not part of the project scope above — the directions worth planning for next.
Get in front of this market — PR that reaches the buyer
MPS sells to a specific, hard-to-reach audience: cleanroom engineers, facility managers, and quality/compliance directors. They don't respond to broad advertising — they trust the sources their peers trust. The highest-leverage marketing move is to become a recognized name in the media this market actually reads, and in the AI tools they now use to make decisions.
Trade & industry PR. Placements, expert commentary, and technical bylines in the publications and associations cleanroom professionals follow — turning the 25-year reputation into visible authority.
Being cited by the AI assistants these buyers ask before they buy. When a compliance director asks an AI "who supplies ISO 5 cleanroom garments," MPS should be in the answer. Earned now, through structured content and authoritative coverage — and early movers are hard to displace.
Ongoing content marketing
- SEO — grow the organic channel that already returns the most revenue per session.
- Email — nurture, reorder, and reactivation on the Phase 2 platform.
- Social & blogging — LinkedIn-led B2B presence and application guides that feed both search and PR.
Paid media, tuned to what converts
- The paid channels already return strongly on cleanrooms and consumables.
- Ongoing management, offline-conversion tracking, and budget reallocation toward what drives revenue.
- Keeps the engine efficient as it scales.
Boundaries & next steps
Not included, or held for later
- Full QuickBooks integration. Phase 1A includes only a discovery session and connection check. Building the live QB → dashboard (or QB → orders) pipeline is its own scope once the requirement is defined.
- PO process simplification. Pulling cost of goods and last-price-update onto the in-house PO — groundwork is in 1A, the workflow is separate.
- Automations build. Phase 2 plans the abandoned-cart and dormant-customer flows; building them out is the next step after the planning session.
- Content & product photography for the rebuilt pages.
- Ongoing maintenance. Keeping the reduced plugin set current and caches healthy — better as a standing arrangement.
To move this forward
- Confirm the pricing workflow. Google Sheet vs. Excel and whether the QuickBooks discovery session is included.
- Confirm the WP Engine plan qualifies for the AI Toolkit, ahead of Phase 3 Options A and D.
- Decide the email platform — Mailchimp or Klaviyo — for Phase 2.
- Pick the AI options to start with; Option A (smart search) is the fastest, lowest-risk win.