Skip to content
bartovsky_
Theme
martin@bartovsky.com

[00] independent erp & bi consultant

Numbers you can trust, from the warehouse to the board report.

ERP, BI and AI for companies that make, stock and sell goods - and whose data doesn't add up yet. I fix how data gets captured, build the reporting on top, and turn it into decisions: why a number changed, and what to do about it.

Get in touch ->See the work ->

$ whoami martin bartovsky · paris area · erp & bi, both sides of the table

trace / gross margin reportdemo
why_margin_is_wrongexample
  1. 01reportgross margin, Q3, all stores38.2%?
  2. 02modelmargin = sales - cost of goodsok✓
  3. 03erpSKUs with cost price = 01,240x
  4. 04sourcegoods received without a priceroot!
  5. 05fixprice required at receipt + nightly checkdone✓
trace / material shortagedemo
why_material_is_missingexample
  1. 01reportassembly stopped, part K-14 missing0 pcs?
  2. 02erporder line: delivery on 12 Octopen✓
  3. 03mailsupplier moved it to 26 Octinboxx
  4. 04sourcedate change never reached the ERProot!
  5. 05fixone route for every date + 5-day checkdone✓
trace / ebit vs budgetdemo
why_ebit_missed_budgetexample
  1. 01reportEBIT year to date vs budget-1.3M?
  2. 02salesrevenue above budget+2.2%✓
  3. 03costmaterial share 44.1% vs 41.5% planned-1.2Mx
  4. 04sourcesteel surcharge not passed onroot!
  5. 05actsurcharge clause with 2 key customersowner✓

[01] services

Three things, done properly.

For companies that make, stock and sell goods - manufacturers, retailers, e-shops.

  • ERP

    ERP projects

    Implementation, migration and consolidation - translating between the business and its suppliers, and asking the awkward questions on both sides.

    • +Process maps in BPMN - where handovers break
    • +Roles, permissions and master data
    • +Migration supervision
  • BI

    BI & AI

    Power BI and Microsoft Fabric models on SAP, SQL Server and Oracle data - with AI on top where it saves real hours.

    • +Semantic models built for decisions
    • +Plain-language questions to the data
    • +Cloud or local models, where data shouldn't leave the company
  • DEC

    Analysis & decision support

    From a number to a decision: why it changed, where in the process it started, what follows - and who acts on it.

    • +Cause and effect, down to the process
    • +Trends from snapshots, not just today
    • +Read-outs with one action and an owner

[02] approach

Capture. Model. Analyse. Decide. In that order.

A report is not the end of the job. A dashboard on top of bad capture shows the wrong number faster - and a right number nobody acts on changes nothing.

  1. 01CAPTURE

    CAPTURE

    Fix how data gets captured

    Map the process first - who does what, where work is handed over. Then master data, booking rules, who may change what.

    erp receipt = qty + price

  2. 02MODEL

    MODEL

    Model it once, for decisions

    One semantic model on the dimensions the business decides by.

    model sales · stock · margin

  3. 03ANALYSE

    ANALYSE

    Find the cause, not a culprit

    Compare over time, split the change, follow it back into the process - with AI where it saves real hours.

    why -8% = price + volume + mix

  4. 04DECIDETRUST ✓

    DECIDE

    Turn it into a decision

    A read-out with one action and an owner, checked against the next snapshot.

    act owner · date · next check

  5. TRUST ✓
receivables risk - illustrative example, synthetic data8 examples
Receivables risk dashboard: aging, weekly snapshots, top overdue customers, comments from finance and a read-out with one action

Every page answers one question and ends with a read-out: what changed, why, and what to do. Eight examples ->

[03] point of view

Most reporting problems aren't reporting problems.

They start where data gets captured. And AI on top of messy data doesn't fix it - it just gets things wrong faster.

Bring me one report nobody trusts ->
symptom -> where it really starts
what you seewhere it starts
Stock that exists only in the systemGoods movements booked late, or not at all
Margins nobody believesCost prices missing or never maintained
E-shop and ERP orders don't matchNo shared key, manual re-entry between systems
Dashboards nobody opensDimensions designed for accounting, not for decisions
Reports everyone reads, nobody acts onNo read-out, no owner, no check against the next snapshot

[04] selected work

Recent projects.

All work ->

[05] stack

The systems your numbers pass through.

  • SAPSAP ERPfinance and production data
  • SQLSQL Serversource data and views
  • ORAOraclesource systems
  • HEHelios iNuvioCzech mid-market ERP
  • QADQADmanufacturing ERP
  • PBIPower BImodels, DAX, reports
  • FABMicrosoft Fabriclakehouse and pipelines
  • WEBE-shoporders and stock sync
  • AILLMscloud or local
  • MCPMCPAI connected to data
  • XLExcelwhere it usually starts
  • +your systemif it stores numbers

[06] about

Between the warehouse and the board report: stock that exists only in the system, margins nobody trusts, dashboards built on dimensions that were wrong from day one. That's where I work.

More about me ->
Based
Paris area, France
Role
ERP, data and decisions - no vendor ties
Business
Martin Bartovsky EI, France
Languages
English, Czech, Polish (passive) - French in progress

[->] next step

Where do your numbers stop adding up?

Start with a short call. Bring one report nobody trusts - I'll trace it back to where the data gets captured.

martin@bartovsky.comLinkedIn ->