The sentence on the brochure
However a dealer frames it, the claim comes out the same: tractor GPS guidance pays for itself, because every pass you stop overlapping is seed, fertiliser, chemical and diesel you stop buying. The mechanism is true. What rarely comes with it is a number measured anywhere near a Kenyan field, and when a number does appear it was usually taken from a boom sprayer on a flat American forage block. That gap is what this piece deals with. In Kenya, tractor GPS guidance is sold on arithmetic the buyer is rarely shown. The honest response is to show that arithmetic, say where its inputs came from, and leave blank what no source fills.
Three blanks remain. The published record gives no kit price for Kenya. It gives no litres-per-acre or kilograms-per-acre saving. It does not establish the acreage below which buying the kit is a mistake. If a grower is quoted all three with confidence, that quote comes from a sales sheet, not a trial. What the record does have: a measured overlap percentage with and without a guidance aid, accuracy figures tied to speed, a description of three equipment tiers, and a clear picture of how small most East African plots are. That is enough to build a decision on. It is not enough to promise a payback, and this piece will not.
Five percent, measured in Georgia with foam markers
A foam marker is a nozzle at the boom tip. It drops blobs of foam so the driver can see where the last pass ended, and it is the cheapest guidance there is. The University of Georgia forage bulletin on GPS guidance options reports that operators using foam markers during chemical application tend to overlap about 5 percent of the application width on each pass. The same bulletin states that a lightbar guidance aid can bring that overlap to under 3 percent without increasing the number of skipped areas. As far as the published record goes, those two figures, 5 percent and under 3 percent, are the whole quantitative basis for the payback claim.
Read what those figures are and are not. They came from chemical application, on forage ground, in the southeastern United States, by operators who already had a foam marker running. They are percentages of implement width, not litres or shillings. The bulletin frames the payoff as more efficient chemical use and fewer total passes, which leads to less time, fuel, chemicals and seed used. It then gives no per-acre figure for any of those four. Its authors had the data to make a dollar claim and chose not to. A Kenyan dealer who makes one has less data than they had.
On a Kenyan block, the baseline changes, not the lightbar. Most operators here do not run foam markers at all. Georgia describes the alternative: driving off the previous tyre track or the line of disturbed soil. The bulletin says that has obvious disadvantages and causes both overlaps and skips. It adds that with a spinner-disc fertiliser spreader or a boomless sprayer, it is very hard to tell where the previous pass actually landed. So a Kenyan starting overlap is probably worse than 5 percent, and nobody has measured it. That is the honest answer. Skips matter too, and the payback sheet does not show them: a strip missed by the spreader is fertiliser you paid for elsewhere and crop that got none, and no overlap percentage captures it. Before you believe a saving, measure your own overlap. On a dry pass, mark the boom tips with a bag of lime, walk the field, and tape the doubled strip. That afternoon's work is the only Kenyan overlap figure that will ever exist for your farm. It is also the first input in the sum below.
The arithmetic on a phone
Here is the template. It uses the Georgia percentages and your own implement. Take your working width in metres. Multiply it by 0.05. That strip is what a foam-marker operator doubles on each pass. Multiply the same width by 0.03. That is the lightbar strip, and because the bulletin says under 3 percent, treat 0.03 as the ceiling. The difference between the two is the width you stop treating twice on every pass. Suppose your boom or spreader pattern is 10 metres wide, and use your own figure instead of this one. Five percent of 10 metres is 0.5 metres. Three percent of 10 metres is 0.3 metres. The saving is 0.2 metres of width per pass, which is 2 percent of everything the implement puts out.
That comes to two percent of the chemical bill, two percent of the fertiliser and two percent of the fuel for that operation. It is also two percent of the seed if the implement is a drill. Now multiply by the number of guided passes you make in a year across every operation that can use the aid. The bulletin lists sprayers, spinner-disc spreaders, seed drills and forage harvesting equipment as the kit. Add those passes up. That annual sum is the only payback number that belongs to you. It depends almost entirely on how much product you push through the machine in a season.
If your true baseline is worse than 5 percent because you drive off tyre tracks, the saving is larger. If you spray a small block twice a season, it is tiny. To turn a per-pass saving into a yes or no on hardware, use the same method the sensor payback piece on this site applies to probes: count the decisions the tool changes in a year, price each one, and compare against the quote. Guidance changes a decision every pass. That sounds like a lot until the passes are counted.
What the saving is made of
Overlap costs you four things, and they don't weigh the same. Chemical and fertiliser is the line the Georgia bulletin was written about: put product on the overlapped strip twice and you have bought it twice. Seed only comes in when the guided implement is a drill. Time and fuel work by another route, fewer total passes, because a field with less overlap gets covered in fewer runs.
Fuel behaves differently. Chemical savings track the overlapped width: treat a strip twice and you pay for it twice. Fuel savings come mostly through pass count. With less overlap, each pass covers more new ground, so the field is finished in fewer runs, and the Georgia bulletin counts fewer total passes as part of the payoff. On a long rectangular block that is simple enough. On a small, irregular Kenyan plot with plenty of headland turns, it isn't. The turns burn diesel whatever the guidance does, and the pass count may not fall at all, because the field is only a handful of passes wide and a narrower overlap doesn't take out a whole pass. The grower who should care about the fuel line is the one on large contiguous ground, running wide booms and many passes a season, already watching the diesel bill. For that grower the Georgia arithmetic transfers most directly. No source gives a litres-per-acre saving. This piece won't either.
Yield isn't on that list. Georgia claims reduced overlap, fewer skips and fewer passes, along with less time, fuel, chemical and seed. It makes no claim that the crop yields more, and you shouldn't either. If a dealer adds a yield line, ask which trial it came from.
Three tiers, and the one that was built for America
The Georgia bulletin describes guidance running from free smartphone apps with low position accuracy all the way to very expensive systems accurate to within an inch, and its advice is to work out first what accuracy the operation needs. Three tiers sit inside that range. First, the phone or tablet, where the internal GPS alone gives precision anywhere from less than 3 ft to plus or minus 30 ft. Second, the lightbar, on the market almost 30 years, typically a differential GPS or augmentation receiver and antenna, a processor, and a row of lights or a screen telling the driver to steer left or right. Third is autosteer, where the system turns the wheel itself.
The FAO's review of adoption economics for digital automated technologies sets out the history plainly. Before autosteer existed, GPS guidance meant manual guidance. Guidance with automatic steering control came in on the American market, designed around that market's field sizes and tractor fleet. Nothing in the record shows autosteer in wide use in Kenya, and this piece won't claim it is.
No source gives a Kenyan price for any tier, so there is no price range here. What can be said is the shape of the choice. A phone costs nothing you haven't already paid. The lightbar is the tier the 5-to-under-3 percent figure was measured on, which makes it the only tier with a measured saving behind it. Autosteer buys inch-level accuracy, and the question is whether any operation on your farm needs an inch. Spraying and spreading don't. Row work might. And the inch turns out to be a speed question too.
Why the free app fails at spreading speed
Most Kenyan operators will try the phone first, and it disappoints for a mechanical reason, not a vague one. Phones and tablets, the Georgia bulletin explains, update their position at roughly 1 Hz, once a second, while most commercial guidance units update at 5 or 10 Hz. At 1 Hz, a tractor spreading fertiliser at 10 mph travels about 15 ft between one position fix and the next. In the bulletin's judgement that produces too much error at higher field speeds. A lightbar's advantage is as much the update rate as the receiver.
That transfers straight to a Kenyan block, because the physics doesn't care where the field is. Slow the tractor and the gap between fixes shrinks. The arithmetic is speed multiplied by one second. An operator spreading on a hired tractor, at whatever speed the contractor prefers, is at the mercy of that gap, and a phone on the dashboard will draw a path that lags the machine. Fix quality is a separate problem. Newer phones use cellular towers to sharpen their position, which is called assisted GPS, while a tablet without a cellular signal falls back on its bare internal receiver, the one with the plus or minus 30 ft worst case.
Then there are failures that have nothing to do with accuracy. Running guidance nonstop eats phone battery fast. You can plug it in, if the tractor has a socket. Some apps have bugs and restart or quit in the middle of a pass. Many won't work without data, so producers are told to check signal before they depend on one. Any Kenyan grower who has lost signal in a valley knows which end of the field that happens. That doesn't make the free app worthless. It makes it a tool for slow passes on small blocks with signal, and for finding out whether you care about guidance at all before you spend.
An accuracy claim means nothing without a speed
Ask a dealer who claims centimetre accuracy what speed that figure is for. In a Danish trial described in the Agris record on an RTK-guided side-shift hoe, the system tracked the crop row to plus or minus 16 mm at P = 0.95 at 0.31 m per second. Push forward speed to 0.52 m per second and the same system held plus or minus 22 mm. Same receiver. More speed, more error.
That trial is not a guidance spec. It used an autonomous tractor and a side-shifting frame, both with RTK-GPS, and a rotor hoe with tines released one by one to miss geo-referenced plants. The plants were artificial, set at a mean within-row spacing of 0.2 m. Accuracy came from a Kalman filter, tilt signal processing and purpose-built control algorithms, not from the receiver alone. Even then, tines entered the protected 10 mm zone around a plant in 18 of 1224 observations. It was a research rig on flat Danish ground. None of those millimetres carry over to a Kenyan hoe.
What does carry over is the shape of the trade-off. Whatever tier you buy, the accuracy you get is the accuracy at your working speed, on your slopes, with your tilt. For spraying and spreading, a lightbar at field speed is the tier with figures behind it. For mechanical weeding between rows, precision costs speed, and on a hired machine speed is paid by the hour.
A cheap fix that only works on straight lines
Some error in any guidance setup is geometry, and geometry costs nothing to fix. A 2011 study from the universities of Valladolid, Burgos and Kentucky, published as a simple method to improve autonomous GPS positioning for tractors, starts by admitting that error is always there when you guide a tractor along a desired path. The usual fixes, better receivers, better differential corrections, or fusing GPS with compasses and inertial units, are complex and expensive. Their alternative: put the receiver ahead of the tractor and use kinematic laws, or a geometric approximation, to work out where the midpoint of the rear axle actually was.
The study reported an effective reduction in guidance error along a straight trajectory. That qualification matters more than the result. Turns and headlands were not tested, and the model assumed front-wheel steering with unsteered, forward-driven rear wheels. On a long straight wheat run, that assumption holds. On a small irregular block with lots of headland turns, which is what most Kenyan fields are, the turns are where the error lives. The method says nothing about them. Practical lesson: ask the installer where the antenna goes and what the unit assumes about the distance from there to the boom tip or drill. That conversation costs nothing and quickly tells you whether the seller understands the kit.
Under two hectares, the question changes
Everything so far assumed you own the tractor and your fields are long enough for overlap to show up as a line item. The European research summary on supporting East African smallholders says farmers in the region usually work plots smaller than two hectares, yet produce up to 70 percent of their country's food. On a plot that size the per-pass arithmetic still works. But there are few passes, the implement is often a knapsack or a hired machine for one day, and a 2 percent saving on a small bill will not reach a kit price in any realistic year. No source gives a break-even acreage and this piece will not invent one. The direction, though, is not in doubt.
For that grower the right call is usually not to buy. A free phone app on a slow pass costs nothing and teaches the overlap habit. Better still, let the hired tractor carry the guidance. The argument in the planter post on this site was to hire mechanised sowing through a group before anyone talks about buying. The same logic holds here. A contractor running many clients' acres across a season piles up the passes that make a lightbar pay, and the client gets the straighter pass in the hire rate. Ask the contractor whether the machine has guidance. If the answer is tyre tracks, Georgia's warning about overlaps and skips is the one that applies.
It is also worth a look at what the cheapest digital spend buys at this scale. In their account of reaching smallholders on feature phones, PlantVillage and iShamba report that 350,000 SMS messages a week costs $1,800. That is advisory messaging, not guidance, and it says nothing about overlap. What it does show is that on small plots your technology budget stretches furthest on information. Hardware comes after.
The pass that overlap cannot rescue
Under the whole payback claim sits a quieter assumption: that the pass being straightened was worth making in the first place. Guidance trims a few percent off a spray pass. A pass made into a drying wind, or onto a canopy still wet past noon, can waste most of the product whatever the boom did. Across 14 NuaSense weather stations in Kenya, over 06 October 2026 to 05 November 2026, 23 percent of 7486 station-hours scored 60 or better on our hourly spray window quality index, and 55 percent scored below 30. Those are aggregated readings from a past month, not a forecast, from stations that are not co-located. Most hours in that record were poor hours to spray. No guidance kit changes that.
So the sequence the crop protection pages here keep coming back to applies: decide when to make the pass, then make it straight. A KES 30,000 weather station computes the spray index, leaf wetness and vapour pressure deficit from its own readings, with the dashboard included. It will not steer a tractor. What it does is tell you which of the straightened passes were worth the diesel, and on a small block that is the larger lever. The NuaSense piece on IoT in Kenyan agriculture covers the low-cost sensor and phone platforms smallholders are using for exactly that timing decision. Put guidance on the large contiguous block with many wide passes and a tractor you own; there, insist on a quote tied to an accuracy at your speed. On a hired day on a small plot, ask what is on the cab, and spend your own money on knowing when to send him. The spray drone against a knapsack comparison reached the same place: the tool earns its cost on passes, and passes are counted, not promised.