Approach

Assess. Prove. Then plan.

Most data methodologies are a sequence of phases with a business case at the front. Ours puts the evidence at the front, because the business case written before the evidence is a guess with a spreadsheet attached.

The four stages

You can stop after any one of them and still be better off.

Each stage produces something that stands on its own. Nothing is engineered so that stopping early wastes what you've already paid for.

Establish ground truth

Stakeholder interviews, system walkthroughs, and profiling run against real extracts. We don't score you off a questionnaire — questionnaires measure what people believe, and the gap between belief and data is usually the finding.

Weeks 1–3

Take the quick wins

Every assessment surfaces two or three fixes worth doing straight away — a duplicate rule, a broken feed, a field nobody has owned since 2019. We do those while the plan is being agreed.

First 30 days

Run the six-month plan

Sequenced by data domain, with a named deliverable every month and a formal checkpoint at month three where you can redirect or stop.

Months 1–6

Hand it over

Runbooks, documentation, training, and a period of supported operation where your team drives and we watch. Ongoing support is a separate decision you make afterwards, not a dependency we designed in.

Closeout
The scale we score against

Where you are, per domain.

The same four states apply to every data domain, and the sequence you move them in is a real decision. Fixing product master data before customer master data can be exactly right — or exactly wrong — depending on which one your revenue runs through.

Unknown Nobody can say what exists Mapped Systems and flows are documented Owned Every domain has a named steward Governed Policy is enforced, not aspirational
01

Unknown

No agreed inventory of what holds the data. Different teams give different answers, and no answer is provably current. Any roadmap written here is fiction.

02

Mapped

Systems, flows and dependencies are documented and agreed. You know where the data lives and how it moves. Nobody is yet accountable for its condition.

03

Owned

Every domain has a named steward with the authority to decide. Issues have somewhere to escalate to, and definitions have someone who can settle them.

04

Governed

Policy is enforced by controls rather than goodwill. Quality is measured continuously, exceptions are worked, and drift is visible before it reaches a report.

Most organisations that describe themselves as governed are at Mapped, with a policy document nobody has read since it was approved. That's not an insult — it's the ordinary result of buying the framework before establishing the inventory.

Why six months

Long enough to change something. Short enough to still be true.

Multi-year data programs fail for an unglamorous reason: the organisation they were designed for stops existing. A restructure, an acquisition, a new finance system, a sponsor who leaves — and month nineteen of a thirty-month plan is now solving a problem the business no longer has.

Six months is roughly the horizon over which an enterprise environment stays stable enough to plan against honestly. It's also short enough that a sponsor can see the whole thing through and be held to the outcome.

If your problem genuinely needs two years, we'll tell you — and then we'll cut it into four blocks, each of which is worth having on its own.

A typical six-month shape
Month 1
Quick wins live
Month 2
First domain modelled
Month 3
Match & merge in test · checkpoint
Month 4
Stewardship running
Month 5
Downstream consumers cut over
Month 6
Handover & measurement

Indicative only. The real sequence comes out of your assessment.

Questions we get asked

The awkward ones, answered.

What if the assessment says our current program is wrong?+
Then that's what the report says, with the evidence for it. It's happened, and it's the single most valuable thing an assessment can produce — a program stopped in month four costs a fraction of the same program abandoned in month twenty. We'll also set out what's salvageable, because it's rarely all of it.
Do we have to use your software?+
No. We built TellParrot because we needed somewhere to hold an assessment that stayed current, and it's how we run ours by default. If you already own a governance or catalog platform, we'll deliver the assessment inside it. The findings are yours in an open format either way.
Can you work alongside our existing systems integrator?+
Yes, and it's a common shape — we handle data domain modelling, quality and governance while an incumbent partner runs platform delivery. It works best when the split is written down at the start. We'll propose that boundary rather than leave it to be discovered in month three.
What do you need from our team?+
For the assessment: access to a nominated contact per major system, read access to representative data extracts, and roughly an hour each from eight to twelve people. For delivery: a decision-maker who can settle definitional disputes, and a steward per domain who has the time the role actually takes. We'll be explicit about that time commitment before you commit.
Where does our data go during the engagement?+
Preferably nowhere. Profiling runs in your environment where that's possible. Where extracts are necessary, we agree the scope in writing, work with the minimum viable sample, host in Australian regions, and destroy them at closeout with confirmation. If your privacy or security team wants that in the engagement letter, we'd expect them to.
How do you price it?+
The assessment is fixed price against a fixed scope. Delivery is priced per six-month block, broken down by month so you can see what you're buying. We don't do open-ended time and materials on the core work — it removes our incentive to be efficient with your money.

Start with the assessment.

Two to three weeks. You get a map of your systems, named owners, a ranked list of gaps, and a costed 6-month plan. No obligation to run it with us.