Mahesh
Mishra

Regional drill-through revealed Q3 underperformance concentrated in 2 territories, not product mix.

EBITDA margin erosion traced to SG&A growing 2× faster than revenue — not revenue slowdown.

Time-to-hire spikes correlated with specific hiring managers, not role scarcity — a process issue.

OTIF failures clustered in 3 suppliers responsible for 60% of volume — Pareto principle confirmed.

Email campaigns delivered 4× higher ROAS than paid social at 1/10th the cost — budget was misallocated.

Automated variance commentary reduced time leaders spent interpreting data by ~3 hours per week.

Dead stock was concentrated in 12% of SKUs consuming 38% of warehouse space — a hidden cost centre.

Top 8% of customers by RFM score generated 52% of revenue — extreme concentration risk identified.

Shift-change periods accounted for 44% of unplanned downtime — a process gap, not equipment failure.

Capex overspend was consistently front-loaded in Q1 — a forecasting bias corrected by revised templates.

North region growth masked by currency effects; constant-currency view revealed true organic growth of 22%.

74% of funnel drop-off happened between consideration and intent stage — not at the checkout.

Sales density gaps mapped to competitor strongholds, revealing whitespace opportunity worth £4M.

Premium category contributed only 11% of volume but 38% of margin — underpriced relative to willingness to pay.

Month-3 was the critical churn point across all cohorts — users who reached month 4 had 85% retention.

3 product pairs had lift > 8× with high support — bundling opportunity not visible in standard reports.

Performance bimodality identified — large middle cluster with no differentiation between B and C players.

Mobile traffic was 61% of sessions but contributed only 22% of conversions — a UX gap, not intent gap.

Logistics costs eroded 6 percentage points of margin on low-ticket SKUs shipped via express courier.

Pricing was 12% above market average for mid-tier SKUs while premium tier was actually underpriced.

Siloed source systems prevented cross-channel attribution; unified warehouse resolved in one design pass.

15% of customers appeared in multiple source systems as duplicates, inflating active customer count.

Manual adjustments to legacy reports introduced errors in 23% of submissions — a systemic risk.

Table scans on 3 large fact tables caused 80% of pipeline latency; clustered index redesign resolved instantly.

Ad-hoc query sprawl caused inconsistent business logic across 6 teams — standardisation was critical.

Missing index warnings in execution plans pointed to 5 high-impact queries generating 70% of server load.

Data quality failures were occurring silently; 11% of daily records had null keys going undetected.

Shadow-run validation revealed 3 stored procedures behaving differently in Azure due to compatibility mode.

Batch inventory updates caused 30-minute blind spots in stock visibility, driving phantom overselling.

Last-touch attribution was over-crediting paid search by 40%; data-driven model revealed email's true role.

Terminal value accounted for 78% of DCF output — model sensitivity to WACC assumptions was extreme.

Manual consolidation took 3 analysts 2 days each month and introduced formula errors in 40% of cycles.

Managers were tracking 30+ KPIs with no prioritisation — cognitive overload was reducing data-driven action.

RAG visibility revealed 60% of delayed projects shared a single resource bottleneck across the portfolio.

Manual commission calculation errors caused payroll disputes costing £30K/year in corrections and goodwill.

Sensitivity revealed project NPV was most sensitive to utilisation rate assumptions, not cost inputs.

ERP export contained inconsistent product codes across regions — causing silent data mismatches for 2 years.

Headcount planning was disconnected from revenue forecasts — leading to over-hiring in 3 consecutive quarters.

Two product lines had negative contribution margins when fulfilment costs were properly allocated.

Forecast accuracy degraded sharply beyond month 4 — seasonal factors were being under-weighted in the model.

Board feedback: previous decks were data-heavy but decision-light — structure was the core problem to solve.

QBR prep consumed 2 analyst-weeks per quarter — mostly reformatting data that already existed in Power BI.

Analysis showed a customer persona previously targeted by marketing didn't actually exist in purchase data.

Investor Q&A consistently challenged TAM assumptions — visual waterfall of TAM → SAM → SOM resolved objections.

Primary research revealed target segment's buying criteria ranked price 4th — feature set was the decision driver.

Stakeholder survey showed previous text-heavy reports had a 23% read-through rate — visuals were essential.

Historical campaign ROI data showed paid video was consistently underperforming against projected returns.

Process mapping revealed 7 redundant approval steps adding 4 days of cycle time with zero risk mitigation value.

Leadership understood model value only when framed as cost saved per decision, not as accuracy percentages.

Stakeholders consumed less than 20% of a 30-slide monthly pack — most slides were never acted upon.

Undocumented models became unmaintainable after the original analyst left — knowledge transfer was the real risk.

Scope creep was the #1 cause of project overruns — absence of a formal BRD was the root cause in 80% of cases.

Inconsistent formatting across team reports was eroding credibility with senior stakeholders — branding mattered.

Absence of data ownership meant critical datasets had no accountable custodian — a compliance exposure.

Engineering and data teams had divergent assumptions about pipeline behaviour — undocumented edge cases were the cause.