Appearance
NanaSelect — Product Vision & Strategy
Last updated: 2026-07-14
Vision
NanaSelect is a standalone product-selection app that guides architects, builders, and homeowners to the right NanaWall system for their project — not the most expensive one that happens to clear the filter. It turns an open-ended, under-informed browsing experience into a guided, educational path that ends in a confident, well-matched, right-sized recommendation. By owning the selection logic and product knowledge — while borrowing the Drupal site's imagery and detail pages over JSON — NanaSelect fills the missing "which system, and why" step that today drives costly mis-specification. It is also a new pathway into the existing nanawall.com/resources configurator (where cost is actually calculated) and into the broader product resources for whatever systems the user selects or compares.
Beyond the selection engine, NanaSelect is becoming the home of NanaWall product expertise itself. The entire nanawall.com content corpus — blog posts, resources, brochures, project case studies, and the linked collateral (PDFs and other files) behind them — is ingested wholesale from the Drupal database and made queryable, powering NanaSage: an authoritative NanaWall expert that grounds guided selection, answers open questions, and shapes the "which NanaWall, and why" narrative customers increasingly receive from AI. Where the app once only borrowed Drupal's imagery and detail pages live over JSON, it now also absorbs Drupal's full body of knowledge as its own corpus — through a single offline ingestion path, so the whole DB is the corpus and there is one source of truth rather than a web of live dependencies.
Program Context
NanaSelect is the second project in the headless nanawall.com program — NanaAwards was the first. The program is a deliberate effort to build new customer-facing capabilities outside the legacy Drupal nanawall.com engine: standalone, modern services that augment the existing site over JSON rather than extending its display-oriented taxonomy and technical debt. NanaSelect owns its own selection logic and product database, and borrows Drupal's strengths (imagery, PDPs, projects, the configurator) through the JSON endpoints — consistent with Principle 4, Standalone, but connected. Framing NanaSelect this way matters for reviewers and new contributors: it is not a Drupal feature, it is a distinct product in a growing headless portfolio.
Problem Statement
Architects frequently start their NanaWall journey by selecting the wrong product — not wrong for the application, but over-engineered and over-cost. A classic example: an architect specifies the NW Acoustical 645 for its high STC rating when a cheaper NW Aluminum 640 — or a far cheaper SL45 — would have satisfied the actual requirement. They see a field like "System STC Required," pick a high number without being educated on what it means, and the current product finder — which filters as you go and leaves many undifferentiated options — never corrects them.
The cost of this is real and downstream. Once an architect specifies an over-spec, over-cost system and moves on, they are very hard to reconnect with. When the project goes out to bid, the contractor value-engineers the expensive system out — swapping it for a cheaper product or a competitor's — and NanaWall loses the sale it appeared to have won. The core failure is upstream in product selection, not in the 3D configurator downstream. That upstream step is the missing piece, and no existing tool owns it well: the Drupal site's product taxonomy was built to power /projects and /resources displays, not to drive an informed selection engine.
Target Users
| Persona | Description | Primary Need |
|---|---|---|
| Specifying Architect / Designer | Professional selecting a system for a client's project; often time-pressed and defaults to over-specifying to be "safe." | To reach a confident, correctly-sized selection quickly — with enough education to defend it at bid. |
| Builder / Contractor | Evaluates or value-engineers a specified system before/at bid. | To confirm the spec is right-sized (so there's no reason to switch it out) or find the appropriate alternative. |
| Homeowner / Residential Buyer | Less technical buyer exploring options for a home project. | More guidance and plain-language education to narrow many systems to a shortlist. |
| NanaWall Technical Advisor / Sales | Internal expert who helps customers select. | A structured tool to guide conversations and receive a warm hand-off at the selection stage. |
Product Principles
Ordered by priority — when principles conflict, higher wins.
- Educate through every decision — the root failure is uninformed selection. Each step should teach the trade-off it's asking about, so the user understands why an option matters, not just picks a number.
- Right-size, don't up-sell — steer to the most cost-appropriate system that genuinely meets the requirements, and actively surface cheaper alternatives that suffice. A right-sized spec is the one that survives value engineering.
- Guide, don't just filter — prefer a linear, decision-driven path that converges on a recommendation over an open filter that leaves a pile of undifferentiated options.
- Standalone, but connected — own the selection logic and product data as a clean, modern service; borrow Drupal's strengths (imagery, PDPs, projects, configurator) over JSON rather than inheriting its technical debt.
- Meet the expert where they are — offer a fast path for professionals who already know the system, without forcing them through the full guided flow, while still nudging them past mis-specification.
What We Won't Do (Non-Goals)
- Not a 3D configurator. Configuration, cost calculation, and CAD/Revit export live in the nanawall.com/resources configurator; NanaSelect ends at the recommendation and is a pathway into that tool (and into other product resources), not a replacement for it.
- Not a replacement for the Drupal website. Product detail pages, resources, and project galleries stay in Drupal; we link to and embed them, we don't rebuild them.
- Not rebuilding the Drupal product taxonomy. NanaSelect maintains its own selection-oriented product database rather than bending the display-oriented taxonomy into a selector.
- Not the cost/quote engine. Actual cost is calculated in the nanawall.com/resources configurator, which NanaSelect routes users into. Within selection we use relative tiers (Budget / Mid / Premium / Ultra) to differentiate systems and inform right-sizing — we don't reproduce pricing or generate binding quotes.
- Not a lead-capture form first. Selection and education come first; contact/advisor hand-off is offered, never a gate.
- Not a general door marketplace. We select among NanaWall systems, not competitors' catalogs.
- Not a standalone DAM (for now). Media is a subsystem serving selection, not a media product; the borrow-from-seebod posture, the fork triggers that would change this, and the integration path if they fire are governed by the DAM spin-off framework.
Strategy
Competitive Landscape
| Alternative | Strengths | Weaknesses | Our Differentiator |
|---|---|---|---|
| Current NanaWall product finder | Already live; covers the full catalog | Filters as-you-go and leaves many undifferentiated options; no education; enables over-spec | Guided, educational path that converges on one right-sized recommendation |
| Drupal 3D configurator | Strong visualization; CAD/Revit/PDF export | Downstream of selection; assumes the product is already (often wrongly) chosen | We own the upstream "which system, and why," then hand off cleanly |
| AI / LLM overviews (ChatGPT, Google Overviews) | Already answering "best NanaWall for X" for customers | Narrative is uncontrolled and may misinform or favor competitors | A structured knowledge base + full content corpus, and a corpus-grounded expert (NanaSage) answering in NanaWall's own voice — which we can also expose to shape third-party narrative |
| Competitor selection tools | Varying guidance quality | Not NanaWall-specific; don't right-size across our range | Deep, NanaWall-specific right-sizing across 24 systems and their real trade-offs |
Key Differentiators
- Right-sizing engine — recommends the most cost-appropriate system that meets stated requirements and explicitly calls out cheaper alternatives that suffice (the 645 → 640 → SL45 insight, encoded).
- Education-first guided flow — a stepped questionnaire built from real selection factors (interior/exterior, geography, climate, residential/commercial, ADA, acoustics, opening shape, material, stacking, sill depth, structural) that teaches as it asks.
- A structured product knowledge base and a full content corpus — the comparison matrix (24 systems × ~60 attributes) as a first-class, queryable dataset, alongside the entire nanawall.com content corpus (blog, resources, brochures, case studies + linked collateral) ingested from Drupal through one path. Together they power recommendation, comparison, an authoritative expert companion (NanaSage), and AI/LLM narrative control.
- Warm, well-timed hand-offs — a deep-linked pathway into the nanawall.com/resources configurator (where cost is calculated) and into the product resources for the selected/compared systems, plus a route to NanaWall technical advisors for those who want a human — each surfaced at the right moment in the flow.
Phased Roadmap
| Phase | Focus | Success Signal |
|---|---|---|
| Now | Shape the data structures from the placeholder CSVs; standalone stepped selector + side-by-side comparison over the product knowledge base; ingest imagery/PDP links from Drupal via JSON; deep-link selected/compared systems into the nanawall.com/resources configurator and resources; encode right-sizing for the acoustic/structural over-spec case. Runs on Cloudflare (Workers + D1 + Pages/R2). | An architect completes a guided session, receives a right-sized recommendation with a cheaper-alternative call-out, and is routed into the configurator to see cost. |
| Next | Full admin UX for NanaWall staff to maintain the product knowledge base (CRUD over systems, attributes, factors, and right-sizing rules); persona-aware paths (pro fast-path vs. homeowner guided); cost-tier differentiation baked into recommendations; explicit "you selected X — Y may suffice for less" alternatives; technical-advisor hand-off; contextual entry points from other site sections (folding, sliding, etc.). Stand up the Product Expertise Corpus — the whole Drupal DB (blog, resources, brochures, case studies + linked files) ingested via one offline path — and the NanaSage expert companion it grounds. | NanaWall owner maintains product data without a developer; measurable shift toward right-sized selections; rising advisor hand-off and configurator completion rates. NanaSage answers real NanaWall questions grounded in the ingested corpus, including linked collateral. |
| Later | AI/LLM narrative control (structured/schema feed for Overviews and chatbots) — now backed by the full ingested corpus, not just the matrix; embedded selection widgets across site contexts; a feedback loop from bid outcomes to tune right-sizing. | NanaWall's own data shapes third-party AI recommendations; reduced value-engineering switch-out at bid. |
Success Metrics
| Metric | Current | Target | Timeframe |
|---|---|---|---|
| Guided sessions ending in a recommendation (vs. filter abandonment) | N/A (filter has no defined endpoint) | Establish baseline, then grow | Post-launch |
| Right-size rate — sessions where a cost-appropriate alternative is surfaced | 0 (no such feature today) | Surfaced in the majority of over-spec-prone sessions | Now → Next |
| Over-specification reduction — recommended tier vs. historical selection tier | Baseline TBD | Downward trend | Next → Later |
| Value-engineering switch-out / competitor loss at bid | Anecdotally high, unmeasured | Reduce | Later |
| Selection-before-configurator/download completion | Unmeasured | Establish and grow | Now → Next |
| Technical-advisor hand-off rate from selection | ~0 from this step | Establish and grow | Next |
Risks & Open Questions
- Data quality. Ownership is resolved — NanaWall will own and maintain the product data via the full admin UX, and the placeholder CSVs are used to shape the data structures for now. The remaining risk is quality and completeness: the POC matrix still has gaps and "??" cells that must be filled before the data can be authoritative.
- Guided vs. filter tension. How hard do we constrain the flow? Too linear frustrates experts; too open recreates today's under-informed filtering. The pro fast-path must still nudge past mis-specification.
- Representing cost without prices. Relative tiers differentiate systems, but will they be enough to change behavior without publishing numbers?
- Drupal integration surface. The app depends on the JSON endpoints (exposed for the nanapad app) for images and project photos — how stable and complete is that coverage? What's the fallback if an endpoint changes?
- The reconnection problem persists. Even a great selection tool can't fully solve the difficulty of re-engaging an architect after they've moved on; right-sizing at selection time is the mitigation, not a cure.
- Can we actually influence AI/LLM narrative? Exposing structured data may help, but third-party Overviews are outside our control — treat as a bet, not a guarantee.
- Adoption. Will architects adopt a new tool over ingrained habits and the existing finder? The path must be demonstrably faster and better.
Capabilities
GitHub Issues labeled cap, grouped by roadmap phase:
Now (p1) — the foundation and core loop:
- [ ] #1 Product Knowledge Base & Data Model — structure the matrix as the source of truth, using the placeholder CSVs to shape the schema
- [ ] #3 Guided Selection Flow — stepped, educational questionnaire from the selection factors
- [ ] #4 Recommendation & Right-Sizing Engine — map answers to systems; surface cost-appropriate alternatives
- [ ] #2 Admin UX — full CRUD for NanaWall staff to own and maintain the product data
Next (p2–p3) — differentiation and reach:
- [ ] #5 Product Comparison View — side-by-side of shortlisted systems
- [ ] #6 Drupal Integration — imagery, PDP links, project photos via the existing JSON endpoints
- [ ] #7 Configurator & Resources Pathway — deep-link selected/compared systems into the configurator (for cost) and resources
- [ ] #8 Persona-Aware Paths — pro fast-path vs. homeowner guided experience
- [ ] #9 Technical-Advisor Hand-off — surface human help at the right moment
- [ ] #47 Product Expertise Corpus — the queryable knowledge corpus behind NanaSage
- [ ] #1094 Drupal DB-dump Bulk Ingestion — bootstrap the full nanawall.com catalog offline
- [ ] #1364 Drupal DB is the Corpus — one DB-dump ingestion path feeds the corpus incl. linked files; retire the JSON:API fetch arm
- [ ] #760 NanaSage — corpus-grounded NanaWall expert chat companion
Later (p4) — narrative control:
- [ ] #10 AI/LLM Narrative Feed — structured/schema output (now corpus-backed) to shape third-party recommendations
This document is the anchor for all product decisions. Capabilities and requirements should trace back to this vision. If something doesn't connect here, question whether it belongs.