Software development proposal template

Ironwood runs 22 concurrent projects across the Pacific Northwest. Every one of them generates daily inspection records: safety, quality, subcontractor handoffs, equipment checkouts. Today those records live in a mix of paper clipboards, iPhone photos in group texts, and a shared Dropbox folder organized by project manager preference. It works. It also loses about a day per superintendent per week, and it's the single biggest source of billing disputes with your subs.

We're proposing a custom field inspection app: an iPhone app for superintendents and foremen to run inspections in the field, and a web dashboard for project managers and the office. Every inspection is timestamped, geotagged, and tied to a project and a sub. Photos are attached in-line. Disputes are resolved against the record, not against memory.

We're proposing a build, not an off-the-shelf deployment, because every mid-market construction platform we evaluated (Procore, Raken, SafetyCulture) either doesn't fit your specific inspection types or costs more at your headcount than building the thing you actually need. The tradeoffs behind that decision are in the Scope section. Read them before you sign.

The data exists, but you can't use it

Your superintendents are running 15–30 inspections a day. They're filling out paper forms, photographing them, emailing them to their PM, and putting the originals in a folder in the truck. When a billing dispute comes up 11 weeks later, the PM has to dig through 3 months of emails and Dropbox folders to find the one photo that proves the sub didn't patch a subfloor before pouring. They usually find it. It takes them 90 minutes. They do this six times a week.

The workflow changes faster than your tools

Every time OSHA updates a safety requirement, or a new insurance carrier requires a new form, or a GC client standardizes on a different subcontractor sign-off sheet, Ironwood has to update a PDF, email it to 34 superintendents, and hope they print the new one instead of using the old stack in their truck. At last count you had 19 active inspection form types, and the rate of change is accelerating.

You've already tried the obvious alternatives

Procore is $42k/year at your seat count, charges per project, and its inspection module is a relatively small feature in a much larger suite you don't need. Raken is cheaper but doesn't support your custom sub-specific signoff workflows. SafetyCulture is close but the photo handling and the offline behavior on rural job sites is unreliable. You've tried two of these and one paid pilot; they each work for 60% of your use case and leave the other 40% to Excel. At your scale, the cost of a 60% solution is bigger than the cost of a 100% solution built to fit.

We build this in four phases, with an explicit list of what ships in each phase and what doesn't. Most custom software projects don't fail on technical problems. They fail because scope grows faster than the team can build, and by month four nobody remembers what was supposed to be in the MVP. Explicit scope fences are the cheapest insurance we know.

Discovery (Weeks 1–2)

Two days on site with two of your superintendents, watching them run inspections end-to-end on real projects. We record workflow timing, photograph the paper forms they use, and interview your PMs about what they need from the office side. The output is a 20-page requirements document, a data model, and a list of ten specific workflow decisions we need answered before week three. This is the only phase where the scope is still open for negotiation; after week two it's locked and changes go through a change-order process.

MVP Build (Weeks 3–9)

We build the iPhone app (React Native, because your superintendents are 80% on iPhone with a long tail of older Androids) and the web dashboard (React, TypeScript, Postgres, Fastify on AWS). The MVP supports: 5 inspection form types you use most (we'll pick these in discovery), offline mode with conflict resolution, photo capture with compression, geotag + timestamp, sub and project assignment, and a PM dashboard with search and filtering. Auth is Google Workspace SSO because that's what you already use. Nothing custom where a standard exists.

Pilot (Weeks 10–12)

We deploy to two active job sites: one big, one small, both within driving distance of Ironwood HQ. Real superintendents use it every day for three weeks while we watch what breaks. We fix every bug and every confusing interaction. By end of pilot the app is cleared for company-wide use on the forms we built.

Launch & Handoff (Weeks 13–14)

Company-wide rollout. Two training sessions (one in Portland, one in Tacoma). Written runbook for your IT lead. The app goes into production with a 30-day warranty period; any bug we find in that window, we fix on our dime. After that, ongoing work moves into a support agreement or you can take the whole codebase in-house. It's yours, with full source and documentation.

Included in the MVP

• iOS app (React Native), built against iPhones running iOS 16+. One Android build included for the two foremen who use Android
• Web dashboard (React + TypeScript) for project managers and office staff
• Backend: Postgres, Fastify, AWS (ECS Fargate + RDS + S3 for photos)
• 5 inspection form types, defined during Discovery (our recommendation: daily safety, subcontractor signoff, quality QA/QC, equipment checkout, and incident report)
• Offline mode with automatic sync when connectivity returns, including conflict resolution for the 1% of cases where two people edit the same inspection
• Photo capture in-line with compression and EXIF metadata preservation
• Google Workspace SSO for all users; no separate login to remember
• Project + subcontractor management (importable from your existing spreadsheets)
• Search and filtering across all inspections by project, sub, date, form type
• CSV export for accounting and insurance
• Admin interface to manage users, projects, subs, and form definitions
• 30-day warranty period after launch for bug fixes at no additional cost
• Source code, documentation, and AWS account credentials handed over on final payment

Explicitly not in the MVP

• Integration with QuickBooks or Sage for automatic billing (designed for phase 2 if the pilot succeeds)
• Automated inspection form builder. Form definitions for v1 are edited by us, not by Ironwood. (This is a deliberate MVP cut. It's the single biggest scope-creep risk.)
• Advanced reporting or analytics dashboard beyond search, filter, and CSV export
• Subcontractor-facing portal. Subs still sign paper or sign on a super's device for v1
• Android parity beyond the one build we've committed to
• PDF form generation for insurance carrier submissions (phase 2)
• Compliance certifications (SOC 2, HIPAA, etc.). The data we're storing doesn't require them, but if your insurance carrier starts asking, that's a separate engagement

Weeks 1–2: Discovery

On-site days in week 1 (one day at an active job site, one day at HQ). Requirements document and locked scope by end of week 2.

Weeks 3–9: MVP Build

Frontend and backend build in parallel. Weekly Monday standup with your PM lead (30 minutes, on Zoom). End-of-week demo video every Friday so you can see progress without needing a meeting. Staging environment available from week 5 onward.

Weeks 10–12: Pilot

Two active job sites using the app daily. We attend each pilot site once a week in person for the first two weeks. We fix every pilot bug within 48 hours.

Weeks 13–14: Launch & Handoff

Company-wide rollout, training sessions, handoff of source code and documentation. Final payment triggers on successful launch.

Weeks 15–18: Warranty Period

30 days of included post-launch bug fixes. After this, any additional work moves to a support agreement or your in-house team.

Threadline Labs is a 6-person custom software studio based in Seattle. We build internal tools and B2B software for mid-market companies: construction, logistics, healthcare, specialty manufacturing. We don't do consumer apps, we don't do venture-stage MVPs, and we turn down about 70% of the projects we're asked to quote because they're not a good fit for our shape.

How we work

Every project has one engineering lead who stays on it from discovery through launch. No handoffs, no offshore subcontracting, no staff augmentation model. You talk to the person building the thing. When we say weeks 3–9 is the MVP build, it's the same two engineers for the whole seven weeks.

Recent work

• Cascade Freight Solutions: dispatcher tool replacing a 15-year-old AS/400 screen. 16 weeks, in production 14 months, zero downtime incidents
• Meridian Anesthesia: pre-op checklist app for surgical teams, HIPAA compliant, deployed to 4 hospitals
• Kestrel Metals: QA/QC platform for a specialty steel fabricator, integrated with their existing ERP

What clients say

"Threadline rebuilt the dispatcher tool our company had run on for 15 years. They hit every milestone, we launched on schedule, and 14 months later we've had zero downtime. The first software vendor we've recommended to other companies in our industry."

Mike Delaney, COO, Cascade Freight Solutions

Payment Schedule

• 20% ($17,000) due on signing. Covers Discovery and kicks off Build
• 30% ($25,500) due at start of Build (Week 3)
• 30% ($25,500) due at start of Pilot (Week 10)
• 20% ($17,000) due on successful launch (end of Week 14)

Invoices are payable within 14 days. Late payment beyond 30 days pauses work; we'd rather have the hard conversation than quietly fall behind schedule.

Change Requests

Scope is locked at end of Week 2 (Discovery). Changes after that point go through a written change-order process: estimate, impact on timeline, impact on cost, your written approval before any work starts. We're not precious about this (most projects have 1–2 changes mid-build), but we are strict about writing them down. Informal changes kill timelines faster than anything else.

Intellectual Property

On final payment, all source code, designs, documentation, and AWS account ownership transfer to Ironwood Construction. There are no ongoing licensing fees, no hidden dependencies on our infrastructure, and nothing stopping you from taking everything in-house or giving it to another vendor. Threadline retains the right to feature the project in our portfolio in generic terms (client name, industry, scope) unless you ask us not to.

Warranty & Post-Launch Support

The 30-day warranty covers bugs: things that worked on launch day and stopped working, or things specified in scope that don't do what they were supposed to. It doesn't cover new features, changes requested after launch, or issues caused by changes you make to the infrastructure. After the warranty, additional work is billed at $180/hour or against a monthly support retainer if you want one.

Cancellation

Either party may cancel with 14 days written notice. Ironwood is billed for all work completed to date and any committed third-party costs. If we cancel, the most recent milestone payment is credited back pro-rata.

Proposal Validity

This proposal is valid for 30 days from the cover date. Engineer capacity is the binding constraint. If you need to defer start by more than 6 weeks, we'll need to re-confirm availability.