Palm oil starts as a heavy red fruit bunch cut from a tree on a family farm. In Indonesia, small farmers working a few hectares each grow more than 40% of the country’s oil palm area, and almost 3 million of them depend on it [1]. They sell their bunches to a local collector, who trucks them to a mill that presses out crude palm oil; the mill sells that oil to refiners and exporters, and in the end food and consumer-goods makers, many of them selling in Europe, pay for it.
Those end buyers now want proof of where each bunch grew. On 12 November 2026, Indonesia’s deputy foreign minister, Arif Havas Oegroseno, told the Indonesian Palm Oil Conference, the industry’s yearly meeting in Bali, that smallholders “dibebani syarat pembuktian yang sangat berat” (are burdened with very heavy proof requirements): data tracing the fruit, the location of the plot and land papers showing the farm did not come from cleared forest. Without them, he said, the product is rejected and financing is closed [1]. The European Union’s deforestation rule applies to large companies from 30 December 2026 [3], and Indonesia’s own sustainability certificate, ISPO, becomes compulsory for smallholders by 2029 [6].
Around the Tesso Nilo national park in Riau province, the proof breaks at the collector. Mongabay Indonesia, an environmental news site, counted about 34 palm oil mills within 50 kilometres of the park, and quoted a collector who buys farmers’ fruit there: “Kita tidak terikat. Ke mana harga lebih dominan tinggi, buang (jual) ke situ” (We are not tied to anyone. Wherever the price is higher, that is where we sell) [15]. A mill buying from such collectors does not know which farms its fruit came from, and tomorrow the same fruit may go to a rival. That gap is the opening for a company that does not exist yet, which this article designs and tests on paper: Shared farmer records for Tesso Nilo mills, one survey team that maps each farm once and sells the record to every mill that buys its fruit.
From a farm to a European shelf
The chain has five links, and only the last two are large companies. Farmers sell their fruit for cash at a collection point, called a peron or ramp, usually owned by a local collector. The collector picks a mill each day by price: “Tiap PKS beda harga. Ada harga tinggi, tapi potongan tinggi juga” (Every mill pays a different price; some pay more but also deduct more), one told Mongabay [15]. The mill weighs the truck, presses the fruit and sells crude oil to a refiner.
How palm fruit travels from a farm to a European buyer
- 1. Family farm
A smallholder on a few hectares cuts fruit bunches and sells them for cash - 2. Collector at a ramp
Buys from many farms and picks a mill each day by price - 3. Mill
Weighs the truck, presses the fruit into crude palm oil - 4. Refiner and exporter
Checks suppliers’ proof and can drop a mill - 5. Food and consumer-goods makers
Pay for the oil; those selling in Europe must show where it grew
The refiner is where the proof is checked: the plot’s location, its land papers and a showing that it was not cleared forest [1]. When Mongabay followed trucks carrying fruit from inside Tesso Nilo, it traced oil to a refinery owned by Apical, one of Indonesia’s large palm oil groups. Apical told Mongabay it had asked the supplying mill, PT Gemilang Sawit Lestari, to explain, and the mill stopped working with the collector concerned [15]. Refiners can drop a mill. What they do not do is map the mill’s farmers for it.
A mill without proof is not shut out of every market. It can sell untraced oil to domestic food, chemical and fuel buyers, but farmers’ groups report that fruit without complete tracing is marked down by “hundreds of thousands of rupiah per ton” [5]. Nationally only about 7% of smallholder palm land is ready for the European rule, and only about 3% carries the STDB, the basic plantation registration issued by district offices [2].
Why nobody maps these farmers
The mechanism is simple. Mapping is paid for by a big buyer, and a big buyer pays only for the farmers who supply its own mills. A mill outside such a programme cannot easily pay for its own mapping either, because the farmers its collectors buy from also sell to rival mills, so each mill would be paying to document fruit that may go to a competitor next week. The farmers who sell through collectors to unsponsored mills are therefore mapped by nobody.
Mapping programmes exist, and they work for the farmers inside them. Wilmar, a large palm oil refiner and trader, gives free mapping and certification help to independent farmers in its supply chain: 4,620 farmers on 10,661 hectares so far. A Wilmar manager said that farmers supplying other mills “whose companies are unable to provide these services” need outside funding to hire consultants [12]. Unilever paid Meridia, a farm-data company, to map 28,000 smallholders in Aceh, North Sumatra and Riau, all in Unilever’s supply base [4]. Koltiva, an Indonesian traceability company, mapped and registered at least 5,000 smallholders for the plantation group REA in East Kalimantan [13]; its own staff describe the cost of such work as relatively high and the private response as not yet at scale [13] [14]. Public money helps too: BPDP, the government’s palm oil fund, pays for group certification work under a 2025 farm ministry regulation [7] [8], but only about 1% of smallholder area is ISPO-certified [13].
| Option for a mill near Tesso Nilo | What the mill gets | What it leaves out |
|---|---|---|
| Refiner or brand programme (Wilmar, Unilever with Meridia, REA with Koltiva) | Free mapping Only if the sponsor buys from the mill | Sponsor’s suppliers only No source shows one at the Tesso Nilo mills |
| Hire its own surveyors | About US$7–8 a farmer Estimated cost of enrolling farmers itself [11] | Farmers it can find Collectors’ farmers also sell to rivals |
| Government help (BPDP, district registration) | Paid certification For organised farmer groups | Slow and partial About 3% of land registered [2] |
| Do nothing | No outlay Sell oil to domestic buyers | Discount and lost buyers Markdowns of hundreds of thousands of rupiah a ton [5] |
We searched in Indonesian and English for a mapping company or programme working with the farmers of the Tesso Nilo mills and found none; that does not prove there is none, and Meridia’s Riau work is reported at province level without naming districts [4]. We also found no record of what any mill pays today to document its suppliers. That is the largest gap in the case.
One farm visit, sold to several mills
The proposed company would work one landscape densely. Here is how one farmer’s record would be made, step by step, under the business as we have designed it:
- A mill near the park signs a yearly contract. The company asks the collectors who deliver to that mill to name the farms behind their fruit, offering them a lower markdown at traced mills as the reason to cooperate.
- Before anyone travels, the company draws the likely plot boundary from satellite images, so a surveyor arrives knowing where to walk.
- A surveyor visits the farm, walks the boundary with a GPS phone, photographs the farmer’s identity card and land papers, and helps fill in the STDB registration the district office issues. The farmer pays nothing.
- The plot is checked against the government’s forest map and the park boundary. A plot inside the park or the forest estate is recorded as such and is not sold as clean.
- The record is linked to the delivery tickets of every mill the farmer’s fruit reaches. Each of those mills pays a yearly fee for that farmer. We assume about Rp 100,000–150,000 (about US$6–8), set just below what a mill would spend enrolling the farmer itself. The company revisits plots each year to keep records current.
The mill gets proof for farmers it could not have found itself, without hiring a field team, at a price no higher than mapping them alone. The farmer gets the plot registration and certification paperwork free, and a buyer that marks down traced fruit less. A second income line comes from the same records. BPDP pays, through farmer groups that apply to it, for the field help a group needs to win ISPO certification, carried out by facilitators, consultants or company partners [7] [8]; the company could be one of those providers. For one Riau farmer group certified in 2021, that help cost about Rp 2.5 million per farmer [9].
The arithmetic that decides it
The business stands or falls on one comparison: what it costs to make a record, against what each mill would spend doing it alone. A 2023 estimate put the cost to a buyer of enrolling one farmer in a traceability system at about US$7–8, roughly Rp 130,000–145,000 [11]: that is what a mill would spend doing it itself, not a price any company charges. Our own estimate for the company, all on assumptions, is about Rp 100,000 a record where land papers are easy to obtain: about Rp 50,000 of surveyor time (a surveyor costing Rp 150,000–200,000 a day and mapping three to four farmers a day), Rp 20,000 of travel, Rp 15,000 of software and data, and Rp 15,000 for the work of signing mills and collectors. Where papers need long follow-up, published costs for legality documents run from almost nothing to about US$11 a farmer [10], which can push a record past Rp 300,000.
So for a single mill the company’s cost is close to the mill’s own, and with hard land papers it is higher: farmer for farmer, it has little room to charge less than a mill doing the work itself. The room comes from selling the same record more than once. If, on average, one and a half mills buy each farmer’s record (an assumption) at Rp 120,000 each (the middle of the assumed fee), a record earns about Rp 180,000 against a cost of about Rp 100,000. If each farmer’s fruit goes to only one mill, or surveyors manage only two farmers a day, the margin disappears. The collector’s “we are not tied to anyone” [15] suggests sharing is real, but nobody has measured how many mills each farmer’s fruit reaches. The mill’s side of the bargain is clear in kind but not in size: traced fruit avoids markdowns of hundreds of thousands of rupiah a ton [5] and keeps the mill on refiners’ supplier lists, but we found no figure for how much fruit a farmer here sells a year, so we cannot turn that into a saving per farmer.
The first market is small and countable: about 34 mills [15]. If 10 to 17 of them sign, each paying for 1,000–3,000 supplier farmers (an assumption) at Rp 100,000–150,000, mill fees come to about Rp 1–8 billion a year (roughly US$55,000–425,000). With BPDP certification work, the company could earn about Rp 2–8 billion a year by its second or third year: enough for a field operation of 15–40 people in one landscape, not a large company. It grows only by repeating the model in neighbouring Riau mill landscapes and then in Jambi or Kalimantan; Riau alone has about 1.61 million hectares of independent smallholder palm [14].
How it could fail
Three parties could each stop it. Mills may simply not pay: they can keep selling untraced oil at home, and no budget for supplier mapping was found. Refiners may refuse records collected by an outside company and shared across rival mills (no refiner has said whether it would accept them), or may decide to extend their own free programmes to the mills that supply them, which would remove the customer overnight. And collectors lose something real when farms are named: the freedom to mix fruit from different sources, including fruit from inside the park that Mongabay traced into the chain [15]. Whether a smaller markdown at traced mills outweighs that for them is unknown.
Land is the other risk. About 62% of smallholder palm land in Indonesia is classified as forest area [5]. If most plots behind the Tesso Nilo collectors turn out to lie inside the forest estate or the park, the company would mostly be documenting fruit that cannot be sold as clean. A rival is also close at hand: Koltiva or Meridia could sell the same shared package once someone shows mills will pay. The advantage is density in one landscape, which makes each extra record cheaper; it can be repeated elsewhere but is not hard to copy.
Could it work? The answer turns on two replies
Possibly, but nobody has yet asked the two questions that decide it. Documented: the mills, the collectors who sell to whichever mill pays most, the buyers’ demand for plot-level proof, the funded programmes that stop at their sponsors’ own suppliers, and a cost estimate for mapping a farmer. Inferred: that farmers’ fruit reaches more than one mill often enough for shared records to be cheaper. Unknown and decisive: whether a mill near Tesso Nilo would pay a yearly fee per farmer, and whether a refiner would count records shared across mills. The cheapest test is a priced written offer to five of the 34 mills, together with a written answer from one refiner’s supplier-compliance team on a sample set of records. Mill owners, refiner compliance staff and Riau survey firms could answer both within weeks.
What has to be true
The business described above stands on these claims, listed in the order we would test them.
- Third-party mills near Tesso Nilo will pay a yearly per-farmer fee for maintained plot records; no record of what they spend today was found
Not yet tested · Mills
Mills risk deductions and losing refiner buyers, but no budget line or payment was found [5] [11]
Would break if: Five of the 34 mills refuse any fee or say refiners will do it for free
Cheapest test: Ask five mill owners what they pay now for supplier records and put a priced written offer (about Rp 100,000 per farmer per year) in front of them - Refiners buying from these mills accept plot records collected by an independent operator and shared across several mills
Not yet tested · Refiners
Refiners already monitor third-party mills, e.g. Apical asked mill GSL to explain [15]
Would break if: A refiner’s supplier team says it accepts only its own or its vendor’s data
Cheapest test: Show one refiner supplier-compliance team a sample record set and ask in writing whether it would count it - Dealers and ramp owners will tell the operator which farms their fruit comes from
Not yet tested · Dealers
Dealers sell to any mill by price and lose the ability to pass on forest-area fruit [15]
Would break if: Most dealers approached refuse to name farms even for a bonus or a lower deduction
Cheapest test: Approach ten dealers supplying two mills and count how many name farms within a month - Each farmer record is bought by more than one mill on average
Our inference from the sources · Mills
A dealer says ‘kita tidak terikat’ and sells wherever the price is higher [15]
Would break if: Delivery records show most farmers’ fruit goes to one mill
Cheapest test: Match a sample of 100 mapped farmers against delivery tickets at three mills - A surveyor can map and document three to four farmers a day, keeping the cost per record under the US$7-8 a mill would spend alone
Not yet tested · The operator
Enrolment is put at US$7-8 per farmer; field productivity is unsourced [11]
Would break if: A two-week field trial averages two farmers a day or fewer
Cheapest test: Run a two-week manual mapping trial with two surveyors in one village - Enough mapped plots lie outside forest-classified land to be sold as clean records
Our inference from the sources · Farmers
About 62% of smallholder palm land nationally is classified as forest area [5]
Would break if: Most plots behind the landscape’s dealers fall inside the forest estate or the park
Cheapest test: Overlay the first 200 mapped plots on the forest-area map - Refiners and sponsors do not extend free mapping to these mills before the operator signs them
Not yet tested · Refiners
Wilmar staff say farmers of other mills need outside funding [12]
Would break if: A refiner announces free mapping for all its supplier mills’ farmers in Riau
Cheapest test: Ask the refiners buying from the first five mills whether such a programme is planned
Research Notes
The sources, calculations and evidence records behind this investigation. Numbered citations in the article link here; ↩ returns to the text.
Sources
1. Wamenlu sebut adanya diskriminasi dalam regulasi sawit global (ANTARA, 12 Nov 2026) ↩
Full text read. Supports: Deputy foreign minister's statement on proof requirements; smallholders' share of palm area and number of growers.
2. APKASINDO Sebut Baru 7% Kebun Sawit Rakyat Siap Hadapi EUDR (Kontan, Oct 2026) ↩
Full text read. Supports: About 7% of smallholder palm land EUDR-ready; STDB coverage about 3%.
3. Jelang Penerapan EUDR, Traceability Sawit Pekebun Swadaya Jadi Kendala (Kontan, 27 Oct 2026; Kemendag statement) ↩
Full text read. Supports: EUDR application date for large operators (30 December 2026); Kemendag statement on smallholder traceability; exchange rate used (Rp 18,100/US$).
4. Unilever and Meridia Partner to Map Smallholders' Plantations (Palm Oil Magazine, 1 Aug 2024) ↩
Full text read. Supports: Meridia's mapping of 28,000 smallholders for Unilever in Aceh, North Sumatra and Riau; province-level only.
5. Indonesia's Smallholder Palm Oil Farmers Face Crossroads Ahead of EUDR Implementation (Palm Oil Magazine, 13 Oct 2025; INDEF/APKASINDO) ↩
Full text read. Supports: Markdowns of 'hundreds of thousands of rupiah per ton' on untraced fruit; about 62% of smallholder land classified as forest area.
6. ISPO Mandatory for Smallholders by 2029, Government Rolls Out Incentives (Palm Oil Magazine, 12 Jul 2025) ↩
Full text read. Supports: ISPO compulsory for smallholders by 2029.
7. Permentan No. 33 Tahun 2025 tentang Sertifikasi Kelapa Sawit Berkelanjutan Indonesia (pasal.id text) ↩
Read in sections (pasal 2 kriteria transparansi (ketertelusuran sumber pemasok tbs); pasal 23; pasal 67-71 (biaya dan bantua [… clipped]). Supports: Traceability of fruit suppliers as an ISPO criterion; BPDP-funded assistance (read in sections: Pasal 2, 23, 67–71).
8. Kejar Sertifikasi ISPO, Petani Butuh Dukungan Nyata (Republika, 30 Apr 2026; BPDP statement) ↩
Full text read. Supports: BPDP funding of ISPO assistance for farmer groups.
9. Analisis Pembiayaan Sertifikasi ISPO oleh Pekebun Kelapa Sawit Swadaya di Kabupaten Rokan Hulu (Jurnal Triton 15(2), Dec 2024) ↩
Full text read. Supports: Cost of ISPO certification for one Rokan Hulu (Riau) farmer group: assistance Rp 2,507,088 per farmer; total Rp 4,759,886.
10. Cost reduction for upscaling voluntary sustainability standards: independent oil palm smallholders in Central Kalimantan (Irawan et al., Frontiers, Aug 2024) ↩
Read in sections (results and discussion text read; table 3 cells not parsed). Supports: Legality-document costs of US$0.23–10.74 per farmer and certification cost structure (text read; Table 3 cells not parsed).
11. Will the EU's deforestation law keep forests standing? (Dialogue Earth via Eco-Business, Jul 2023) ↩
Full text read. Supports: Estimate that enrolling one farmer would cost about US$7–8 (PISAgro, reported in 2023).
12. Technical knowledge remains the biggest hurdle for independent palm oil smallholders to comply with EUDR (Eco-Business, 6 Nov 2025; Wilmar staff) ↩
Full text read. Supports: Wilmar's free support for 4,620 farmers on 10,661 ha; Wilmar staff on farmers of other mills needing outside funding.
13. Only 1% certified: Tackling barriers to EUDR compliance for Indonesian smallholders (AgNavigator, 11 Aug 2025; REA/Koltiva) ↩
Full text read. Supports: Koltiva's work for REA (at least 5,000 smallholders mapped and registered in East Kalimantan); about 1% of smallholder area ISPO-certified; cost described as relatively high.
14. EUDR: Collaboration vital as deadline looms over Indonesia's smallholders (AgNavigator, 17 Nov 2025; Koltiva co-founder) ↩
Full text read. Supports: Riau's 1.61 million ha of independent smallholder palm; Koltiva co-founder on the private response not yet being at scale.
15. Segel Satgas tak bendung, sawit ilegal Tesso Nilo terus ngalir ke pabrik (Mongabay Indonesia, 30 Apr 2026) ↩
Full text read. Supports: About 34 mills within 50 km of Tesso Nilo (GFW); collectors' quotes; Apical's response on mill PT Gemilang Sawit Lestari.
Competitors and substitutes
| Name | Offering and price | Who it serves | At the Tesso Nilo mills? |
|---|---|---|---|
| Koltiva | Mapping and registration Run for a sponsor; price not published [13] | Plantation groups, refiners, brands e.g. REA in East Kalimantan | Unknown Two Riau-aimed searches found nothing either way |
| Meridia (for Unilever) | Plot mapping and consent Paid by Unilever [4] | Unilever supply base 28,000 smallholders, three provinces | Unknown Riau reported at province level only |
| Wilmar programmes | Free mapping and certification Cost not disclosed [12] | Its own supply chain 4,620 farmers engaged | Unknown No district list published |
| Refiners’ supplier monitoring (e.g. Apical) | Checks and sanctions No fee; burden on the mill [15] | Third-party supplier mills | Yes Apical questioned PT Gemilang Sawit Lestari, within 50 km of the park |
| BPDP-funded ISPO assistance | Group certification help About Rp 2.5 million per farmer in one Rokan Hulu group [9] | Organised farmer groups [7] [8] | Unknown No provider named for these farmers |
| Government systems (e-STDB, SI-ISPO) | Public registers No charge found [3] | All growers STDB about 3% of land [2] | National Hold data; do no field mapping |
| Mill’s own staff | In-house enrolment About US$7–8 a farmer [11] | Larger mills | Possible Reaches only farmers the mill knows |
| Do nothing | Sell untraced oil at home Markdowns of hundreds of thousands of rupiah a ton [5] | Most unsponsored mills | Yes The default |
Economics workings
All rupiah figures use Rp 18,100 per US$ [3]. Rows marked “assumed” have no source and are the ones to test.
| Step | Calculation | Status |
|---|---|---|
| Mill doing it alone | US$7–8 × Rp 18,100 ≈ Rp 127,000–145,000 per farmer [11] | Reported estimate A cost, not a price charged |
| Surveyor time | Rp 150,000–200,000 a day ÷ 3–4 farmers a day ≈ Rp 40,000–60,000 | Assumed |
| Travel | Rp 15,000–30,000 per farmer | Assumed |
| Software and data | Rp 10,000–20,000 per farmer a year | Assumed |
| Signing mills and collectors | Rp 10,000–20,000 per farmer | Assumed |
| Land papers | US$0.23–10.74 ≈ Rp 4,000–194,000 per farmer [10] | Sourced Central Kalimantan study |
| Cost per record | Rp 79,000–324,000; central case about Rp 100,000 where papers are easy | Estimate |
| Revenue per record | 1.5 mills × Rp 120,000 = Rp 180,000 | Assumed Fee and sharing both untested |
| Breaking point | 2 farmers a day and no sharing: cost ≈ Rp 150,000 before land papers, above the mill’s own US$7–8 | Estimate Weakest inputs |
| First market, mill fees | 10–17 of 34 mills [15] × 1,000–3,000 farmers × Rp 100,000–150,000 ≈ Rp 1.0–7.7 billion a year | Mixed Mill count sourced; the rest assumed |
| Cross-check from farmers | Riau 1.61 million ha [14] ÷ 2–3 ha ≈ 540,000–800,000 farmers × 10–15% near these mills × 40% on legal land × 30–50% signed ≈ 6,500–24,000 farmers; × Rp 150,000 ≈ Rp 1.0–3.6 billion | Mixed Area sourced; shares assumed; 62% forest-classed nationally [5] |
| With BPDP assistance | Rp 2.5 million per farmer [9] for about 20% of farmers over 4 years, added to mill fees ≈ Rp 2–8 billion a year in years 2–3 | Estimate |
| Year 5, after expansion | Repeated in two or three more Riau landscapes, then Jambi ≈ Rp 9–47 billion a year | Assumption No comparable entrant found |
Routes considered
Folded into Shared farmer records for Tesso Nilo mills: collectors naming farms at the delivery point (Dealer-Disclosed Delivery Pass, Ramp Operator as Verifier, Dealer Harvest ID Collection, Buyer-Transporter Role Split); linking records to delivery tickets (Delivery-Ticket Data Capture); drawing plots from satellite images first (Satellite-First Plot Registry); the shared per-farmer fee (Shared-Cost Landscape Fee); lighter re-checks for reliable collectors (Dynamic Verification Intensity); collecting STDB papers on the same visit (STDB Dual-Purpose Certification); and reusing the records for BPDP-funded group certification (Pre-Packed Group ISPO Pipeline, Field-data reuse for BPDP group certification). Multi-mill data pooling for shared certification is the same business under another name.
- BPDP-funded group traceability pool: records kept for a group rather than each plot would not meet the plot-level proof buyers ask for.
- Price-Guarantee Forward Contracting: guaranteeing farmers a fruit price needs large working capital and still produces no records.
- Cooperative formation and BPDP application processing: slow public-programme work with no mill paying for it.
- Pre-approval Onboarding and Cascading Refiner Acceptance: both depend on refiners’ own internal rules, which a new company cannot sell.
- Mill team dual-purpose ISPO service: a mill using its own staff is an internal improvement, not a company.
- Two-phase cooperative certification pipeline and Data-quality-weighted group eligibility service: the first may not fit the 2025 certification regulation [7], and the second is a certifier’s rule rather than a service.
Open questions
- What, if anything, do mills near Tesso Nilo spend today on supplier records? Mill owners and their sustainability managers could say.
- Would a refiner’s supplier-compliance team accept records made by an independent operator and shared across mills?
- How many mills does a typical farmer’s fruit reach in a year? Delivery tickets at three mills would show it.
- What share of the plots behind these collectors lies outside the forest estate and the park?
- Do Koltiva, Meridia or Wilmar already work with any of the 34 mills?
Leave a Reply