06 / Data analysis & work management
Start with the records. End with a decision.
At Merchok Agrotrade, I used Excel and SQL to analyse trade activity, sales and operations. I also use Power BI and Tableau for visualisation. My approach is to check the data first, define the measure and show what the result means for the next decision. I use Trello and Odoo to track deliverables and keep owners and next steps clear.
01 / SpreadsheetsExcel
I structure worksheets around a business question and give each row a clear unit, such as one invoice, payment or order. I use lookup functions to join related records by a stable reference and flag missing codes or duplicate keys before I trust the totals.
I use VLOOKUP for exact matches in existing workbooks and INDEX/MATCH when the result column can sit on either side of the lookup key. The XLOOKUP example below shows how to flag a missing product code for review.
For reconciliations I compare source records by reference, date and amount, then keep an exception list with the evidence needed for review. PivotTables and conditional totals help me summarise sales and spending by product, supplier or month. A reviewer can trace the report back to the underlying row.
For a recurring workbook, I separate the imported data, lookup tables and report outputs. A review tab tracks missing references, duplicates and total differences so the person using the report can see what needs attention before refreshing it.
See an example lookup and conditional total
=XLOOKUP(A2, Products[Code], Products[Crop], "Check code")
=SUMIFS(Ledger[Amount], Ledger[Supplier], B2, Ledger[Status], "Open")
Here are two ways to run the same product lookup. The VLOOKUP formula uses adjacent Code and Crop columns, in that order.
=IFNA(VLOOKUP(A2, Products[[Code]:[Crop]], 2, FALSE), "Check code")
=IFNA(INDEX(Products[Crop], MATCH(A2, Products[Code], 0)), "Check code")
SQL
I use SQL to answer trade and sales questions when the source data sits in more than one table. I identify the record keys, select the columns the question needs and join the records at the correct level of detail. I check row counts before and after a join so a duplicate match does not inflate the totals.
I use filters to define the reporting period and grouping to compare categories. A left join is useful when an unmatched record needs to remain visible. I check a sample back against the spreadsheet or export so the resulting count makes sense before it reaches management.
For the crop example, the report grain is one order and the product lookup must have one row per product code. I check for repeat keys before joining, then compare the source order count with the query result and investigate unmatched codes separately.
See the query behind the crop example
SELECT COALESCE(p.crop, 'Unmatched') AS crop,
COUNT(*) AS orders
FROM weekly_orders AS o
LEFT JOIN products AS p ON p.code = o.product_code
GROUP BY COALESCE(p.crop, 'Unmatched')
ORDER BY orders DESC;
Power BI
I use Power BI to turn checked records into a view that leadership can scan quickly. In a service review, I would define the denominator for resolution rate, separate open cases from overdue ones and check that the dashboard totals match the source export.
Then I lay out the headline measures, trend and breakdown so a reader can move from “what changed?” to “which team or queue needs attention?” Filters make it possible to view a period or team without changing the definition of the metric.
For a review of service backlog, I would include the reporting date and age bands, let the reader filter to a team, and reconcile the filtered totals to the underlying list. That gives managers a route from the dashboard to the cases they can follow up on.
See an illustrative DAX measure
Total Cases = COUNTROWS(Cases)
Resolved Cases =
CALCULATE([Total Cases], Cases[Status] = "Resolved")
Resolution Rate = DIVIDE([Resolved Cases], [Total Cases])
In the example, the export contains one row per case.
Tableau
I use Tableau to explore how a result changes across a period or category and turn that into a clear visual. A ranked bar chart can show which crops account for the most orders. A weekly trend can show whether a change looks persistent or came from a single busy period.
I keep the default view focused on the decision. I add clear labels and useful filters and place details in tooltips. If a chart shows workload, I show the volumes and the time window so a manager can decide where to look first.
For a trade report, I would pair ranked product volumes with a weekly trend and make the reporting window visible. I would give users enough context to distinguish a demand change from a missing or delayed source file.
05 / Workspace rolloutTrello
I introduced Trello at Adela Technologies, created the workspace and managed it. I moved action tracking from chat threads to shared boards so colleagues across 6+ team roles could see who owned each deliverable and when it was due.
When a meeting generates several actions, I give each deliverable its own card with an owner, a deadline and enough context for someone to pick it up. If work depends on another person’s input or approval, I record what is outstanding and who has the next action. That gives teams a clear handover as work moves into review.
Before leadership reviews, I check for overdue cards, missing owners and pending approvals. I follow up on the tasks that need attention and bring the CEO the decisions that will move work forward. The rollout improved cross-departmental execution efficiency by 40%.
See an illustrative campaign handover in Trello
A customer education campaign needs copy, artwork and approval by Friday. I create a card for the launch pack, assign an owner, link the draft and add a checklist for those three steps.
When the artwork is ready, I mark that step complete and record who still needs to approve the copy. I close the card when the approved pack has been handed over.
Odoo
At Infytel Communications, I manage the Odoo setup the team was already using when I joined. I work within the existing project and task structure to track commitments from leadership meetings and partner discussions.
I keep the task owner, deadline and status current and record what has changed since the last review. When work is waiting for a technical input, a handover or approval, I note what is outstanding and follow up with the person who can resolve it. If an agreed deadline moves, the task record shows the revised date and next action.
Before a review with the MD, I check open work by stage, owner and due date. I handle routine follow-ups and bring forward delays that need a decision, with the impact on the next commitment. I confirm what was delivered before closing the task so the next review starts with an accurate list.
See an illustrative partner follow-up in Odoo
After a partner call, the team agrees to prepare a scope note before the next meeting. In Odoo, I record the internal owner, the review deadline and the technical input still due.
If that input is delayed, I update the task and tell the MD which commitment is at risk and what decision is needed. I close the task once the reviewed note is ready to share.
