Skip to content

How it works

Two weeks of our work. About four hours of yours.

We follow an order through every system it touches, find where nothing is watching, and design the first fix. Nothing to prepare, nothing installed.

The engagement

From the first email to one journey automated.

Five stages. The assessment is the first four and stands on its own. The fifth only happens if you decide, with the findings in hand, that it is worth it.

  1. Frame 00Before the clock

    You email

    What lands

    A scope, a fixed price, and a straight answer, including no.

  2. Frame 01Day 1–3

    Map

    What lands

    Your critical journeys mapped, with every dependency under them.

  3. Frame 02Day 3–8

    Design

    What lands

    A verdict on each failure, and the right-sized stack to see it.

  4. Frame 03Day 10

    Deliver

    What lands

    The assessment document and a build plan, yours to keep.

  5. Frame 04Weeks 3–8

    Build and run

    What lands

    One journey watched and automated, with the time to notice measured.

Assessment turnaround: 10 working days from the first walkthrough, not from the day you enquire.

Your time

4 hours

That is the whole ask. Just over two of those hours is your people walking us through a real order, which is the part that finds the systems nobody remembers depending on.

An order's path through every system, annotated during a mapping session
5 min

Fill in the form

What you sell, what you run on, and the last thing that broke without warning.

135 min

Walk us through one real order

45 minutes each with two or three people who run the store. Screen shared: click to delivered, and the last few times it went wrong.

30 min

Answer a few questions by email

Usually about volumes, third parties, who gets called first and who is allowed to approve a change.

60 min

Walk through the findings

One call. Where you are blind, what we would fix first, and what we would never automate.

Right-sizing

The smallest stack that can actually see your failures.

Three depths, one methodology. The same business signals and the same detect, investigate and recover lifecycle come out of all three; what changes is how much of your estate is instrumented and what it costs to run. The tool is the last decision, not the first.

01Shopify, Stripe, a shipping API

Lightweight

Business events, synthetic checks and uptime, watching the journeys rather than the servers.

How deep it instruments

30

Enough to know within a minute that orders have stopped, and roughly where. For most businesses this size it is the whole answer, and it is cheap to run.

02A custom checkout or backend

Hybrid

The lightweight layer, plus real instrumentation on the parts you actually built yourselves.

How deep it instruments

70

Where most growing businesses land. Your own code is traced properly; the platforms you rent stay at event level, because that is all you can see of them anyway.

03Custom apps, APIs, Kubernetes

Full

OpenTelemetry across the estate: metrics, logs, traces and APM, correlated into one picture.

How deep it instruments

100

Justified once your own services outnumber your rented ones. It costs real money to run, and we will tell you when you are not there yet.

Guardrails

It will be wrong sometimes. That is designed for, not hoped against.

Every response starts on the first rung. It moves up only once you have watched it be right.

See what each stage costs

01 - It watches and investigates.

Collects the evidence and proposes a cause. Changes nothing.

02 - It acts on your approval.

Shows the action, what it touches and how to undo it. You approve; it runs, then checks orders recovered.

03 - It runs unattended.

Reversible actions only. Never an order, a payment or a customer.

Start with the blind spots

Find out what could break today without anyone noticing.