Farm Labour Bottlenecks: How to Plan a Season Around the Peak Weeks
To plan around farm labour bottlenecks, build a week-by-week labour calendar. Then find the two to four weeks when the work required exceeds the labour you can actually get, and fix those weeks first. In most smallholder systems, those weeks are crop establishment combined with first weeding, and harvest.
The main point of this guide is that the goal is not to save labour across the season. The goal is to shrink or move the single worst week, because a missed window usually costs far more in yield than the wages needed to meet it.
Why farm labour bottlenecks sit in a few weeks, not the whole season
Total labour use for the season tells you little. A household can have spare hands for most of the year and still lose a large share of its harvest because two jobs fell in the same fortnight.
Field data support this. A 2019 study in Agricultural Systems by Silva, Baudron, Reidsma and Giller surveyed about 100 farms in a maize-based system near Hawassa and about 100 in a wheat-based system near Asella, in southern Ethiopia. Their labour calendars showed competition for labour at Asella, but not at Hawassa. The two systems were in the same region, but different crop calendars produced different constraints.
A 2019 CIMMYT-led analysis in Agronomy for Sustainable Development went further. Using farm-level labour data from eight sites across four countries in Eastern and Southern Africa, it found that labour or other farm power limited land productivity at every site. The authors concluded that demand for mechanisation was much higher than macroeconomic studies had suggested, and that the real barrier was access rather than demand.
Missing a window is expensive. CIMMYT-linked monitoring in Mangwende, Zimbabwe, during the 1987–88 and 1988–89 seasons found that the earliest-planted quarter of maize fields averaged 4.9 t/ha, compared with only 1.1 t/ha for the latest-planted quarter. In the same area, 42% of farmers first weeded their earliest maize more than 30 days after emergence, a delay the researchers calculated cut grain yield by 28%.
How much work does one labour day do?
One person-day of hand work covers roughly 0.02–0.05 ha of weeding, basin digging or rice harvesting. A pair of oxen, a herbicide sprayer or a small machine covers ten to fifty times more. The table converts published rates into area covered per working day. Treat the figures as planning ranges, because soil, crop, weed pressure and skill can double or halve them.
| Operation | Hand | Animal traction | Mechanised or chemical |
| Land preparation / establishment | Hand-hoe planting basins: up to 30 labour days/ha in the first year (≈0.03 ha/day) | Pair of oxen ploughing: 9–28 hours/ha, with animals working about 3–6 hours a day (≈0.1–0.7 ha/day) | Rice transplanting: about 1 person-day/ha by machine, against 25–30 by hand |
| Weeding | Hand weeding: about 52 workdays/ha in maize, 54 in sorghum, 33–48 in cassava (≈0.02 ha/day) | Ox-drawn weeding rates: about 7 hours/ha per pair (≈0.4–0.9 ha/day) | Herbicide-based control: mostly under 2 workdays/ha (>0.5 ha/day) |
| Harvest (rice) | Manual cutting: about 160 person-hours/ha (≈0.05 ha per 8-hour day) | Rarely used for cutting | Small reaper: about 7.7 person-hours/ha; mini-combine: about 0.39 ha per machine-hour |
Sources: a 2013 hand-hoe tillage study from Zimbabwe in Agriculture, Ecosystems & Environment (citing Mazvimavi et al., 2008); FAO guidance on draught animal power; IRRI's Rice Knowledge Bank; a 2026 study in Agrochemicals from Ghana (citing Moyo et al.); a 2017 reaper evaluation from Bangladesh published by CIGR; and a mini-combine field evaluation.
Two patterns stand out. In cereal systems, hand weeding is the largest single labour block that remains unmechanised. In rice, transplanting and harvest carry the peaks.
Where do the peak demand windows fall in the cropping calendar?
Peak labour demand falls where establishment, first weeding and harvest of one crop overlap with work on another. Three patterns recur.
Unimodal rainfed maize: southern Africa
Planting follows the first reliable rains, and first weeding falls due within about a month of emergence. That means the early fields need weeding while the last fields are still being planted. This is the overlap behind the Mangwende losses above. Harvest is usually more flexible, because maize can dry standing in the field for a period.
Highland and bimodal systems: East Africa
Two shorter seasons create two smaller peaks, and which one binds depends on the crop. The 2019 Ethiopian study found competition in the wheat system but not the maize system. Build labour calendars per farming system, not per country or region.
Rice-based systems: South Asia
Transplanting alone can require 25–50 person-days per hectare. In Bangladesh, labour charges rise sharply at rice harvest in both the Boro (January–May) and Aman (July–November) seasons, and at wheat harvest. In north-west India, the tightest window is the short gap between rice harvest and wheat sowing, when residue must be managed.
Sequencing and staggered planting: widening the bottleneck without machines
Sequencing costs little cash, but it has a yield price, so calculate that price before you commit.
Staggered planting splits establishment into blocks sown several days apart, so the first-weeding windows overlap less. The cost is lower yield on the later blocks. A 2026 review in Frontiers in Agronomy reports that in south-western Nigeria, fitted losses were 30–38 kg/ha for each day of delay after the first March planting (Fakorede, 1985). At that rate, a 10-day stagger costs about 300–380 kg/ha on the later block.
Priority rules matter as much as dates. When first weeding of early fields competes with planting the last field, the early field usually deserves the labour. It has the higher yield potential, and every day of weed competition there reduces it.
Changing the establishment method removes a peak entirely. PhilRice reports that direct seeding can save up to 23 person-days per hectare compared with transplanting. The trade-off is that direct-seeded rice faces more weed competition, so part of the saved labour moves into weed control.
Selective mechanisation: buy capacity only for the peak
Mechanise the operation that creates the peak, not the whole calendar. For most small farms, that means hiring a service rather than buying a machine.
- Herbicide for first weeding. Tamru and colleagues (2017, European Journal of Development Research, with IFPRI) found that herbicide use on Ethiopian cereals doubled to more than a quarter of cereal area between 2004 and 2014, with labour productivity gains of 9–18% for teff. Adoption was strongest near cities, near all-weather roads and where rural wages were higher. The costs are the product, a sprayer, training and protective equipment. An integrated approach also works: combining herbicide with one hand weeding cut labour by more than 60% compared with weeding twice by hand.
- Ox-drawn weeding. FAO notes that many farmers use draught animals mainly for land preparation and rarely take advantage of their high work rates for weeding. This requires row planting at establishment and a weeding implement.
- Hired harvest. A trial in Barisal, Bangladesh, during the 2018 Aman harvest found that combine harvesting reduced costs by about 58% and labour by 70% compared with manual harvesting.
- Residue-to-sowing turnaround. Shyamsundar and colleagues (2019, Science) estimated that Happy Seeder systems could raise profit by INR 6,000–11,500 (about USD 85–160) per hectare for the average farmer in north-west India.
- Small-scale service models. Baudron and colleagues (2015, Food Security) examined two-wheel tractors, conservation agriculture and private-sector service provision as a mechanisation route for Eastern and Southern Africa.
Worked example: 2 ha of maize and three workers
Assume a household in sub-humid southern Africa with 2 ha of maize and three adults. Each adult works 5 days a week, so the household has 15 person-days per week. Hand weeding takes about 52 person-days/ha. First weeding should finish within roughly 3 weeks, before the 30-day point linked to a 28% loss. That gives 45 person-days available.
Option A: plant all 2 ha in one week. First weeding needs 104 person-days, against 45 available. About 1.1 ha is weeded late. At a 5 t/ha yield potential, a 28% loss on 1.1 ha costs about 1.5 t of grain.
Option B: stagger by 10 days. The two weeding windows now span about 4.5 weeks, or about 66 person-days, but demand is still 104. The shortfall falls to about 38 person-days. At the Nigerian delay rate, the later hectare also loses 300–380 kg. Staggering alone does not close the gap.
Option C: herbicide on 1 ha, hand weeding on the other. Demand drops to about 54 person-days (under 2 + 52). The remaining gap is about 9 person-days, which you can cover by hiring labour.
The decision rule: herbicide pays on a hectare when its full cost (product plus sprayer hire) is less than about 50 × the local daily wage. That comparison does not yet count the grain it protects. Enter your local wage, herbicide price and maize price to check.
FAQ
How many labour days does it take to weed one hectare by hand?
Published figures for smallholder cereals run at roughly 50 person-days per hectare per weeding round, with about 52 for maize and 54 for sorghum in southern African data. Cassava needs 33–48. Heavy weed pressure, late weeding and compacted soil push these numbers higher, so measure one sample plot on your own farm.
What is a labour calendar in farming?
A labour calendar lists, week by week, the person-days each crop operation needs and compares them with the labour the household can mobilise, including hired help. Its purpose is to reveal the specific weeks where demand exceeds supply. Those weeks, not seasonal totals, should drive decisions on planting dates and mechanisation.
Should a small farm buy or hire machinery for peak weeks?
For most smallholdings, hiring is the better choice, because a machine bought for two peak weeks sits idle for the rest of the year. Research across Eastern and Southern Africa suggests the main barrier is access to services rather than demand. Owning makes sense only when a machine can also earn income from custom hire.
Valora Earth's WhatsApp advisor can turn a farmer's planting dates and crop mix into a week-by-week task schedule, so overlapping weeding windows show up before planting rather than after.
