Skip to content

Agree the plan. Build in phases.

Start with a no-charge call. Before each phase, we agree what you get, the price and the schedule.

Packages

Start with a package

Four packages, each with its price on the page. Pick one and buy it online.

Billing
  • Onshore

    US engineers, Central time

  • Senior engineers

    Scope it, then build it

  • Weekly demos

    Working software, every week

  • Price agreed

    Before each phase

One project, start to finish.

An example, not a client: a distributor moves dispatch off a spreadsheet.

  1. 1

    45 minutes · no charge

    First call

    We walk through a day of dispatch and find where it breaks.

    You get

    • A straight answer on what to do next
    • A fixed discovery price, if it fits
    Two people at a table with a laptop showing a spreadsheet and a notebook.
  2. 2

    1–4 weeks typical · fixed fee

    Discovery

    An engineer shadows the dispatchers, maps how a load really moves and prototypes the riskiest routing rule.

    You get

    • A map of how the app will work
    • A plan for what ships first
    An engineer and a dispatcher at a whiteboard workflow diagram in a warehouse office.
  3. 3

    4–6 weeks typical · scope dependent

    First release

    Dispatchers use the first version on dock tablets. Weekly demos decide what is built next.

    You get

    • A working app carrying real loads
    • Tests on every critical step
    A dispatcher in a safety vest checks the dispatch app on a rugged tablet while a forklift loads a truck at a sunlit dock.
  4. 4

    Ongoing, by agreement

    Improve and grow

    Real use shows where trucks lose time, so each month the roadmap follows the data.

    You get

    • A monthly report on what shipped
    • A roadmap they set and can change
    Two dispatchers and an engineer lean over a screen showing delivery routes and a bar chart, with trucks at the docks behind them.
  5. 5

    Ongoing, for as long as you want it

    Maintenance & support

    Patching, monitoring and small fixes continue. The distributor’s team can take over whenever it chooses.

    You get

    • A named person, with an agreed response time
    • Docs, accounts and access in their hands
    In the evening, a laptop on the dispatch desk shows green status lights while trucks sit parked at the docks outside.
Every phase in full
  1. Phase 1

    First call

    45 minutes · no charge

    What breaks today, what that is costing you, and which of our teams would fix it.

    What happens, what you get and how to leave

    We discuss your current process, the problem, your goals and any deadlines or constraints. You leave with a recommendation and, if discovery is useful, a fixed-fee proposal.

    What happens

    • Walk through the current process and where it fails
    • Establish what a good outcome would be worth
    • Identify hard constraints: compliance, deadlines, existing vendors
    • Agree whether discovery is justified

    What you get

    • A direct recommendation on the next step
    • A discovery proposal with a fixed fee, if it’s warranted

    ExitYou leave with a recommendation. There is nothing to cancel.

  2. Phase 2

    Discovery

    1–4 weeks typical · fixed fee

    Paid, fixed-scope, and it ends in a plan you could hand to another firm, plus working code proving the riskiest assumption.

    What happens, what you get and how to leave

    We map the process, systems, data and integrations, then test the highest-risk assumption. You receive a proposed architecture, release plan, estimate and risk register. The service page gives the expected duration; the proposal confirms the schedule and fee.

    What happens

    • Process observation with the people doing the work
    • Domain and data modeling
    • Integration and constraint mapping across the existing estate
    • Technical spike on the highest-risk assumption
    • Interface prototypes tested with real users
    • Sequencing into independently valuable releases

    What you get

    • Solution architecture and domain model
    • Release plan with independently shippable slices
    • Estimate range with the assumptions behind it stated
    • Working spike code and a clickable prototype
    • Risk register with mitigations

    ExitA fixed-fee engagement with a defined end. Continuing is a separate decision.

  3. Phase 3

    First release

    4–6 weeks typical · scope dependent

    A small senior team puts something real to work: a live system, a pilot running on the floor, or a finding you can repeat.

    What happens, what you get and how to leave

    We build an agreed first release and demonstrate working progress each week. Your team can review the code and test the system. We agree changes before adding them to the scope.

    What happens

    • Weekly demos of working software, not slides
    • Continuous deployment to a real environment from week one
    • Automated tests written alongside features, not after
    • Your engineers in the codebase and in code review from the start
    • Architecture decisions recorded as they’re made

    What you get

    • Production system carrying real work
    • Infrastructure as code and reproducible environments
    • Test suite covering the critical paths
    • Architecture decision records

    ExitThirty days’ notice after the first release, for any reason.

  4. Phase 4

    Improve and grow

    Ongoing, by agreement

    The system grows from how people actually use it, not from a plan written before anyone did.

    What happens, what you get and how to leave

    After launch, usage and feedback guide the next release. We review performance, running costs, reliability and the roadmap with your team.

    What happens

    • Usage instrumentation on the flows that matter
    • Performance and cost tuning against real traffic
    • SLOs, alerting and on-call runbooks
    • Roadmap review every month, changeable every sprint
    • Security review and dependency hygiene on a schedule

    What you get

    • Monthly delivery and reliability reporting
    • SLOs with dashboards and an alert policy
    • A roadmap you set and can change

    ExitMonth to month. No minimum term after the first release.

  5. Phase 5

    Maintenance & support

    Ongoing, for as long as you want it

    We keep the system healthy after launch, and your team can take it over whenever you’d rather run it yourselves.

    What happens, what you get and how to leave

    Ongoing support can include patching, upgrades, monitoring and small changes. We agree the coverage and response window. Documentation and handover planning allow your team to take over.

    What happens

    • Dependency, platform and security patching on a schedule
    • Monitoring, alerting and incident response within an agreed window
    • Small changes and enhancements as the business shifts
    • Runbook and architecture walkthroughs, and paired work with your team
    • Transfer of every credential, account and domain whenever you ask

    What you get

    • A named support contact and an agreed response window
    • Complete documentation and runbooks, kept current
    • All accounts and access in your control

    ExitNo minimum term. Thirty days’ notice, for any reason. Every account is already in your name, and your data is returned and then destroyed as the Leaving terms set out.

Choose how we work together.

  • Fixed-fee discovery

    Starting anything meaningful

    • 1–4 weeks, depending on how much estate there is to map
    • Fixed and agreed in advance, against your budget and requirements
    • Deliverables defined in the engagement agreement
    • No obligation to continue
  • Dedicated team

    Sustained delivery against a roadmap

    • Monthly, after the first release
    • Named individuals, not a resource pool
    • Thirty days’ notice, any reason
    • Scale up or down at month boundaries
  • Outcome-based project

    Well-specified scope with a hard date

    • Fixed price against a defined scope
    • Milestone-based payment
    • Change control by written agreement
    • Only offered post-discovery
  • Embedded engineers

    Reinforcing a capable in-house team

    • Monthly per engineer
    • Your backlog, your ceremonies
    • Our review and standards behind them
    • Thirty days’ notice

We don’t bill by the hour

Discovery has a fixed fee. Delivery is priced as a monthly team or a defined outcome. The scope, fee and change process are agreed before each phase starts.

After launch, on purpose

After launch, you can retain a delivery team, agree a support arrangement or bring the work in-house. We set reliability targets for the system and plan the handover in advance.

How we estimate the work.

See the steps and what the estimate contains
  1. Estimate slices, not features

    A feature list contains no delivery risk, so estimating one produces a number that cannot be wrong until it is very wrong. We break the work into slices that are each independently deliverable, and estimate those. It is slower, and it is the reason a range narrows as the project proceeds instead of quietly widening.

  2. Give a range, and say what sits at each end

    You get a low and a high, plus the specific conditions that put the work at either end, not a single number with a padding factor buried in it. If we cannot describe what would have to be true for the top of the range, we do not understand the work well enough to have estimated it.

  3. Write our assumptions as your obligations, with dates

    Every estimate rests on things we assume about your side: an environment exists, someone can approve, a system has a test instance, a vendor answers. Those are listed as dated obligations in the statement of work rather than left implicit, because an unstated assumption is how a contractor manufactures a change order and calls it your fault.

  4. Attach a confidence, and say what would raise it

    Ranges carry a confidence level and the reason it is not higher. Usually the answer is one unknown that can be resolved by building something small, which is why discovery includes a spike against the single riskiest assumption rather than a longer document.

  5. Agree who pays for a miss before the work starts

    Modest variance is ours to absorb; that is the normal play in any estimate. A large miss means the estimate was a guess dressed as an estimate, which is our failure and not a change in your scope. The threshold between those two is a number in the statement of work, agreed before anyone starts, not a conversation in week nine.

  6. Show the last estimate against what it actually cost

    At every phase boundary you get the previous estimate next to the actual, and an explanation of any gap. This is the part firms leave out, and it is the only one that makes the rest verifiable: a method you cannot audit is a claim about our character, and you have no reason to take our word for that yet.

What the estimate contains

  • A low and a high figure, with the conditions that produce each
  • The slices behind them, priced individually
  • Our assumptions restated as dated obligations on your side
  • A confidence level, and the one unknown that would most improve it
  • The overrun threshold and who carries each side of it
  • Enough detail that another firm could price the same scope from it

Terms to review before work starts.

Ownership and licensing, Acceptance and payment, Scope and change, People, Data and security, Leaving
  • Ownership and licensing

    “Work made for hire” is a copyright concept. It says nothing about patents or trade secrets, and it does not reach work done by a contractor, which is why the mechanism matters more than the label.

    • Ownership and licensing settled in writing in the engagement agreement, before any work starts
    • Whatever is agreed is agreed by express assignment rather than by relying on work-made-for-hire language
    • Every engineer and subcontractor under written assignment before they touch the codebase, so there is a clean chain of title to assign
    • Any third-party or open-source component identified, with its license and any fee attached
    • No feedback clause. What you tell us about your business stays yours
  • Acceptance and payment

    Acceptance testing is the protection most services contracts leave out, and the one that decides what happens when a deliverable is not what you asked for.

    • Objective acceptance criteria written into the statement of work, not decided afterward
    • Nothing is “deemed accepted” because time passed
    • Rework to meet the agreed criteria is at our cost, inside a defined window
    • On a fixed-scope phase, payment tied to milestones and acceptance rather than to the calendar
    • Invoices are net 30, paid by purchase card, ACH or purchase order
    • A holdback on each phase: a share of the fee we don’t invoice until you accept that phase
  • Scope and change

    Underbidding a project and recovering the difference through change orders is a known pattern. The defense is a specific scope and a change process you control.

    • A statement of work in active voice, naming who does what and by when
    • Our assumptions rewritten as your obligations, with dates, so neither side discovers them late
    • Every change priced and signed before the work starts
    • Estimates given as a range with the assumptions stated, and a shared overrun mechanism agreed in advance
    • No legal terms buried in a statement of work
  • People

    The named team on the proposal and the team that arrives are frequently different. Staff turnover is the most reliable predictor of cost and delay on a long engagement.

    • Key personnel named in the statement of work and not substituted without your consent
    • Interview anyone you want to before they start
    • You can require removal of anyone, for any lawful reason
    • No charge for the ramp-up time of a replacement we initiated
    • All delivery performed in the United States. Nothing subcontracted offshore
  • Data and security

    Most privacy law holds you responsible for a breach whether it was your fault or your vendor’s. The contract is where that gets allocated.

    • A security requirements exhibit specific to your data, not a pointer to our then-current policy
    • Written notice of any actual or suspected incident within 24 hours
    • You control the content, timing and method of any notification to your customers or regulators
    • Log files and forensic cooperation preserved and provided
    • A signed BAA where PHI is involved, and joint security testing where you want it
  • Leaving

    The cost of leaving is set on the day you sign, not on the day you go.

    • Termination for convenience on thirty days’ notice after the first release, no fee
    • A hand-off phase in the plan from the start, not negotiated under pressure at the end
    • Your data returned in a platform-neutral format as well as ours, then certified destroyed to NIST SP 800-88
    • Documentation, runbooks and architecture decision records written as the work happens
    • A named handover contact for a defined window after the last day, so the last question does not go unanswered

Cost, timing and project questions.

  • Yes, and it is usually the fastest way to get a useful answer: several of the pieces we publish end by asking for exactly that. Two things worth knowing first. This is an ordinary web form: it is encrypted in transit and the message lands in our mailbox and our own console, but it is not a secure document channel, so redact freely. A workbook with the numbers changed, a clause with the parties removed, a day of messages with the identifiers swapped: each of those answers the question as well as the original does, and we would rather read a redacted one. And if you would rather have an agreement in place before you send anything, say so in a line: we work under your paper rather than asking you to work under ours, and we redline rather than refuse.

  • Custom work is quoted against your budget and what you actually need, rather than printed as a range that would be wrong for most of the people reading it. Discovery is a fixed fee, agreed in advance. Delivery is quoted against a written scope, in phases, with each phase’s cost and exit criteria fixed before that phase starts, so the decision to continue is yours at every boundary and there is no open meter. Work that is not a custom build has no minimum: the shortest piece of work on this site took an hour, and the one beside it took four weeks. Tell us the budget you have on the first call and we will tell you what is achievable inside it, including when the honest answer is that it is not. If you would rather know before you spend a call, put the budget you have in your first message and we will answer that question in the reply. The four packages are the exception: each shows its price on its page.

  • On résumé, they are ahead, and you should discount anyone who tells you otherwise. There are onshore firms in this space that have been doing it for eighteen, thirty, even fifty years, with senior engineers and named clients in your industry. The difference we would ask you to weigh is structural rather than technical. Most firms our competitors’ size specialize in one layer and subcontract or decline the rest, so the seams between the control network, the application, the estate and the security review end up owned by whoever is left in the room. We run those as four divisions inside one company, under one contract, which is why the escalation path ends somewhere useful instead of at a vendor boundary. That is the trade on offer: longer history and more logos on one side, and on the other a smaller senior team that can follow the problem across every layer it touches.

Next step

Tell us what’s breaking.

Forty-five minutes, no charge. We’ll say what we’d do, what it would likely cost, and whether what you already have can be made to work.

Reply
A person replies, not a sequence: within one business day, from someone who would be on the engagement.