# Satlink Hotspot Financial Report Reporting period: September 3, 2025 through April 7, 2026, normalized to GMT+3 / East Africa Time. ## Executive Summary Satlink recorded **KES 163,540** in timestamped, deduplicated gross sales during the reporting period across **4,889 paid transactions**, with an average transaction value of **KES 33**. Estimated operating expenses were **KES 40,400**, producing estimated net profit of **KES 123,140** and an estimated net margin of **75.3%**. The strongest month was **2026-03** at **KES 31,580**, while the lowest active month was **2025-09** at **KES 2,160**. The figures are directionally strong but not perfect accounting records. **46.9%** of reported revenue is high-confidence because it has transaction IDs. The remaining revenue is legacy timestamped activity without transaction IDs. There are also **492 undated paid-looking legacy transactions** worth up to **KES 26,220** if every one occurred inside this period; they are preserved in the clean dataset but excluded from period totals because assigning them to months would overstate precision. Operationally, the dataset shows a major evolution from manual/legacy recording into transaction-ID-backed callbacks. Voucher swaps were frequent during the maintenance period and were excluded from revenue. Complimentary or unclassified unavailable vouchers remain visible as data-quality risk, not sales. ## Cleaning Methodology - All tabs were exported from the Google Sheet into `analysis/raw/` and processed from that fixed snapshot. - Columns were normalized to voucher, availability, phone, amount, status, timestamp, transaction id, and notes. - Timestamps were normalized to GMT+3. Timestamped rows without transaction IDs were treated as UTC and shifted forward by 3 hours. Rows with transaction IDs were treated as already GMT+3. - Revenue was deduplicated at transaction level. Transaction ID was the primary key; legacy rows were grouped by phone, amount, status, minute-level timestamp, and sheet. - Rows marked `swap` were excluded from revenue. - Undated paid-looking rows were retained and quantified, but excluded from September 3, 2025 to April 7, 2026 period totals. - Expenses were estimated using KES 4,000 flat per month for seven ISP months, September 2025 through March 2026, and KES 400 per week for electricity prorated for partial weeks. ## Data Audit Summary | Metric | Count | |---|---:| | Raw voucher rows | 6,299 | | Rows retained in cleaned row dataset | 6,299 | | Revenue candidate rows before dedupe | 5,775 | | Duplicate or multi-voucher rows merged | 143 | | Swap rows excluded from revenue | 192 | | Available inventory rows excluded | 145 | | Complimentary/unclassified unavailable rows | 0 | | Paid transactions retained for period analysis | 4,889 | ## Revenue Analysis - Gross sales volume: **KES 163,540** - Estimated ISP expense: **KES 28,000** - Estimated electricity expense: **KES 12,400** - Estimated total expenses: **KES 40,400** - Estimated net profit: **KES 123,140** - Estimated net margin: **75.3%** - Paid transactions: **4,889** - Average transaction value: **KES 33** - Revenue vouchers issued: **4,980** - High-confidence revenue: **KES 76,640** - Legacy medium-confidence revenue: **KES 86,900** ## Monthly Performance | month | revenue | transactions | vouchers | high_confidence_transactions | isp_expense | electricity_expense | total_expenses | estimated_net | avg_transaction_value | mom_revenue_growth_pct | mom_transaction_growth_pct | |:--------|----------:|---------------:|-----------:|-------------------------------:|--------------:|----------------------:|-----------------:|----------------:|------------------------:|-------------------------:|-----------------------------:| | 2025-09 | 2160 | 72 | 72 | 0 | 4000 | 1600 | 5600 | -3440 | 30 | 0 | 0 | | 2025-10 | 25600 | 707 | 715 | 0 | 4000 | 1771.43 | 5771.43 | 19828.6 | 36.21 | 1085.2 | 881.9 | | 2025-11 | 25740 | 782 | 793 | 0 | 4000 | 1714.29 | 5714.29 | 20025.7 | 32.92 | 0.5 | 10.6 | | 2025-12 | 19570 | 603 | 642 | 0 | 4000 | 1771.43 | 5771.43 | 13798.6 | 32.45 | -24 | -22.9 | | 2026-01 | 30540 | 923 | 938 | 579 | 4000 | 1771.43 | 5771.43 | 24768.6 | 33.09 | 56.1 | 53.1 | | 2026-02 | 24840 | 752 | 762 | 752 | 4000 | 1600 | 5600 | 19240 | 33.03 | -18.7 | -18.5 | | 2026-03 | 31580 | 933 | 940 | 930 | 4000 | 1771.43 | 5771.43 | 25808.6 | 33.85 | 27.1 | 24.1 | | 2026-04 | 3510 | 117 | 118 | 117 | 0 | 400 | 400 | 3110 | 30 | -88.9 | -87.5 | ## Weekly Performance Top weeks: | week_start | revenue | transactions | vouchers | electricity_expense | estimated_net_after_electricity | |:-------------|----------:|---------------:|-----------:|----------------------:|----------------------------------:| | 2026-01-12 | 9450 | 239 | 244 | 400 | 9050 | | 2026-03-16 | 7560 | 214 | 217 | 400 | 7160 | | 2026-01-19 | 7200 | 240 | 243 | 400 | 6800 | | 2026-02-02 | 7170 | 220 | 224 | 400 | 6770 | | 2026-03-23 | 6990 | 214 | 215 | 400 | 6590 | Lowest active weeks: | week_start | revenue | transactions | vouchers | electricity_expense | estimated_net_after_electricity | |:-------------|----------:|---------------:|-----------:|----------------------:|----------------------------------:| | 2026-04-06 | 750 | 25 | 25 | 114.29 | 635.71 | | 2025-09-22 | 960 | 32 | 32 | 400 | 560 | | 2025-12-29 | 2730 | 91 | 92 | 400 | 2330 | | 2025-12-22 | 2880 | 96 | 104 | 400 | 2480 | | 2025-12-15 | 3170 | 100 | 106 | 400 | 2770 | Full weekly table is available at `analysis/processed/weekly_performance.csv`. ## Operational Observations - Manual and legacy operation is visible through missing timestamps, missing transaction IDs, and unavailable voucher rows without payment metadata. - Swap rows are concentrated in later operational maintenance windows and should remain excluded from revenue. - The most important reporting improvement is enforcing immutable transaction IDs, callback timestamps in EAT/ISO-8601, and a separate event type field for sale, free issue, manual fulfilment, and swap. ## Recommendations 1. Make M-Pesa receipt/transaction ID a required unique key for all paid voucher allocations. 2. Store all timestamps as ISO-8601 with timezone, preferably UTC plus a display conversion to EAT. 3. Split operational events into explicit types: paid sale, manual paid fulfilment, complimentary voucher, swap, inventory import, and void/refund. 4. Add a reconciliation sheet or D1 export comparing M-Pesa receipts, voucher rows, and SMS delivery. 5. Keep complimentary vouchers in a separate workflow with issuer, reason, and approval fields.