Peaks & ValleysConsulting

Data quality and reporting

The failures that matter are the quiet ones. A formula that has been reading the wrong column since someone inserted one. A query that returns nothing and looks exactly like a real answer of zero.

We reconcile what you maintain against what is true, report the drift, and flag the ambiguous cases instead of silently resolving them.

01 / The problem

A wrong number that looks right is worse than no number.

Every long lived spreadsheet drifts. Someone inserts a column and forty formulas quietly start reading the wrong field. A heading is reworded and a category flag defaults to one value for a year. A lookup fails and returns zero, which is indistinguishable from a real zero.

No one notices, because the report still runs and the number still looks plausible. Decisions get made on it. It surfaces eventually, and usually in front of someone whose opinion matters.

A number you can't rebuild from the transaction data is a number you can't defend. We build reports from the bottom up, so every figure traces line by line back to what happened.

02 / What we do

Everything this covers.

Data quality auditing

  • Reconciling a maintained record against an authoritative source
  • Reporting the drift, with the ambiguous cases flagged instead of silently resolved
  • Finding formulas that have been reading the wrong column since someone inserted one
  • Catching queries that return an empty result that looks like a real answer of zero
  • Validation at every step instead of a check at the end

Record consolidation

  • Duplicates resolved instead of merged blindly
  • One entity, one record, with the history preserved
  • At risk records assigned to named owners instead of to no one
  • The matching rules written down and reviewable, so the result can be audited

Receivables and collections

  • Billing cadence and payment terms
  • The aged book worked down and kept down
  • Unapplied credits found and applied
  • Write off and tax recovery analysis
  • Automation for the long tail of small accounts that no one has time to chase

Accounting systems

  • Migration between accounting systems
  • Integration into the operational platform, so every booking carries its billing contact
  • Approval gates on invoice billed clients
  • Reconciliation of fees, credits and pass through charges

Reporting people can defend

  • Investor and board reporting built from the transaction data up
  • Results against plan, period by period
  • Segment scorecards with their own targets
  • Top clients, geography and pricing
  • Pricing discipline, with every unit below target flagged

Open and external data

  • Automated extraction from public data services, with paging, retries and validation
  • External data joined to your records, with differences reported instead of overwriting what staff entered
  • Scheduled refreshes that flag what changed and why
03 / How it works

The order we do it in.

01

Pick the authority

Decide what the true source is before comparing anything. Half of all data disputes are two systems both believing they're the record.

02

Compare, don't overwrite

Differences delivered side by side for review. Your staff entered that value for a reason, and you should see it before it disappears.

03

Explain the drift

Not just what differs, but why. A column heading that was reworded is a completely different problem from a value someone typed wrong, and they need different fixes.

04

Rebuild from the bottom

Reports constructed from transaction data, so any figure can be defended line by line when someone asks where it came from.

04 / Fit

Who this is for.

A good fit

  • Companies reporting to investors, a board or a lender
  • Anyone whose key numbers come out of a long lived spreadsheet
  • Businesses with duplicate customer records no one trusts
  • Operations reconciling against an external or authoritative register
  • Companies changing accounting systems
  • Anyone who has been surprised by their own report in front of other people
05 / Questions

What people ask about this.

How do you know which version is right?
We don't assume. The first step is agreeing what the authoritative source is for each field, which is a business decision instead of a technical one. After that, differences are reported instead of resolved. Where the answer is ambiguous, it gets flagged as ambiguous instead of being quietly decided by whichever system happened to be read second.
Will you change our data?
Not without you seeing it. Updates to a live record are presented beside the existing value for review, and anything that replaces existing content is backed up first. A reconciliation delivers a side by side comparison, not an overwrite. You decide what gets applied.
Our spreadsheet is a genuine mess. Is it too far gone?
Almost certainly not. Malformed rows, headers that aren't on the first row, dates entered five different ways, numbers stored as text, merged cells: that's the normal condition of a working file, and the tools are built to expect it. A messy file that people use is far more valuable than a clean one no one maintains.
06 / Related

The work next door.

These almost always come up together. An audit that finds a process problem usually finds a data problem underneath it.

Start here

Start with the audit.

Every business is different, so we scope the audit that way. We spend time inside your operation and watch the work actually happen. You get a written map of every valley, what it costs you a year, and what it would take to flatten it. That document is yours whether or not we build anything.

Emailadaignault@peaksnvalleysconsulting.com
Phone(250) 423-9511
FoundersKamila Domanska · Alex Daignault
Based inNelson, British Columbia
ServingCanada and the United States