Every engagement starts at the metric layer. Even the ones that do not want to.
Three practices, seventeen service lines. We scope the work, quote against a written deliverable, hand over, and leave you able to maintain it without us.
One definition per number, written down, owned by a named person, and enforced in a semantic layer instead of a spreadsheet.
Metric audit
The problem: Three reports, three revenue figures, and a meeting where nobody can say which one is right.
A definition sheet your finance team signs off, one row per metric:
Metric name, plain-language definition and formula
Named business owner and the decision it drives
Source of record and the systems that feed it
Reconciliation test against the finance system, with the variance written down
Engagement
Fixed-scope diagnostic
Duration
5 working days
Delivered by
Finance specialist + analytics lead
What we commit to: You get a reconciled number and a written gap list. If the audit finds nothing, we tell you that in writing and you keep the sheet.
Semantic layer build
The problem: Every new question means a new SQL query, and every analyst writes the margin calculation slightly differently.
A governed metric layer that reporting and AI both read from:
Dimensional model with documented grain and relationships
Metric definitions as version-controlled code, not dashboard formulas
Row-level security mapped to your actual org structure
Test suite: uniqueness, referential integrity, and metric reconciliation
Engagement
Fixed-price project in milestones
Duration
3 to 6 weeks
Delivered by
Analytics and BI lead
What we commit to: Week-two demo on real data. If the layer cannot reproduce last year-end’s finance numbers, the build does not leave discovery.
Metric ownership & governance
The problem: Nobody knows who is allowed to change a definition, so definitions drift quietly until a board meeting exposes it.
The operating rules that keep the layer honest after we leave:
Metric council charter, cadence and decision rights
Change process with versioning and a published changelog
Definition review checklist tied to your closing cycle
Handover pack with recorded walkthroughs
Engagement
Fixed-scope workshop series
Duration
2 to 3 weeks
Delivered by
Engagement lead + analytics lead
What we commit to: You leave able to run this without us, which is the point.
02
Data platform
Pipelines that fail loudly instead of silently, a warehouse that answers a query before the meeting ends, and lineage you can hand to an auditor.
Warehouse & modelling
The problem: Your data lives in an ERP, a CRM, four spreadsheets and someone’s mailbox, and no two extracts agree.
A cloud warehouse built to be maintained by someone who is not us:
Ingestion and modelling layers with documented contracts
Incremental loads, partitioning and cost guardrails
dbt (or equivalent) project with tests and CI
Environment separation: dev, staging, production
Engagement
Fixed-price project in milestones
Duration
4 to 8 weeks
Delivered by
Data platform engineer
What we commit to: Query cost and runtime baselines documented before handover, so your finance team can see the bill coming.
Integration & pipelines
The problem: Reports break every time someone edits the export format, and nobody notices for three weeks.
Pipelines with alerting and a runbook:
Connector setup for ERP, CRM, payroll, marketing and POS
Freshness and volume checks with alert routing
Failure runbook with owner, escalation and recovery steps
Backfill procedure you can run yourself
Engagement
Fixed-price or time and materials
Duration
3 to 8 weeks
Delivered by
Data platform engineer
What we commit to: A silent failure is a defect. Every pipeline ships with a test that fails loudly when the source changes shape.
Data contracts & quality
The problem: An upstream team changes a column, and every downstream number is wrong before anyone notices.
Contracts between the teams that produce data and the teams that depend on it:
Schema and semantic contracts, versioned in the repository
Quality monitors on the metrics that drive decisions
Lineage from source column to published metric
Quarterly review pack with breach history
Engagement
Retainer or fixed-scope project
Duration
3 to 5 weeks
Delivered by
Data platform engineer + analytics lead
What we commit to: You find out about a broken source from a monitor, not from a customer.
03
Reporting & enablement
Fewer reports, read more often. Reporting that survives the person who built it leaving.
Executive & operational dashboards
The problem: You have dashboards. The Monday meeting still starts with someone updating a spreadsheet.
Reporting designed around decisions, one page per decision owner:
Decision-first page design with a written brief per page
Thresholds, alerting and narrative commentary instead of raw tables
Performance budget: page interactive in under two seconds on the standard laptop
Adoption baseline and a 30-day usage review
Engagement
Fixed-price project
Duration
3 to 5 weeks
Delivered by
Analytics and BI lead
What we commit to: If a report is not opened in 30 days, we delete it. Fewer pages, used more often.
Planning & decision support
The problem: Forecasts are rebuilt by hand every month from whatever the last spreadsheet said.
A repeatable planning cycle with a driver-based model:
Driver tree linking operational inputs to financial outcomes
Scenario switching for volume, price and cost assumptions
Variance-to-plan reporting with attribution
Monthly cycle runbook mapped to your closing calendar
Engagement
Fixed-price project
Duration
4 to 6 weeks
Delivered by
Finance specialist
What we commit to: Plan versus actual tells you which driver moved, not just that it did.
Team enablement
The problem: The tool is licensed, the training was a webinar, and everyone still asks the analyst.
Enablement measured by tickets deflected, not by attendance:
Role-based training on the premises people actually use
Documented definition wiki with search that works
Champions programme with a named owner per team
Ticket-volume baseline and a 60-day re-measure
Engagement
Fixed-scope programme
Duration
2 to 4 weeks
Delivered by
Analytics lead + engagement lead
What we commit to: We measure inbound requests before and after. If the curve does not bend, we do not bill the last phase.
Engagement models
Three ways to buy, and what each one is for
Which model fits is part of the scoping conversation. The fixed quote follows the written scope, usually within two working days.
Model
How it is priced
Use it when
Fixed-scope project
Quoted against a written scope, fixed before we start
Most work. One deliverable, one date, one number. Change the scope and we re-quote in writing first.
Readiness sprint
Fixed quote after a short scoping step
A week to a fortnight, ending in a decision rather than a commitment. Cheapest way to find out whether the rest is worth buying.
Managed metrics retainer
Monthly, days per month agreed in advance
Two days a month of senior time: definition changes, new pages, pipeline upkeep, and an owner for the model.
Prices exclude VAT and travel. Larger builds are split into milestones so you are never funding more than one gate ahead of delivery.
Not sure which of these you need?
That is what the first call is for. We will tell you which service fits, which one is premature, and whether a cheaper fix gets you most of the way. One new engagement starts at a time, so we are seldom motivated to invent work.