Horticulture for Growth (H4G) Project Promoting Horticulture Enterprise Development in Kenya: Endline Evaluation Report
Farmers report high recall of H4G training and widespread use of improved seeds, fertilisers and crop protection. However, the share growing the project's target crops fell compared with baseline, per-acre production costs rose sharply, and middlemen still dominate sales.
Overview
This endline study evaluated the three-year Horticulture for Growth (H4G) project, led by Solidaridad in partnership with Syngenta. It targeted smallholder horticulture farmers in seven Kenyan counties: Kirinyaga, Makueni, Laikipia, Bomet, Elgeyo Marakwet, Nakuru and Nyandarua. Target crops were tomatoes, Irish potatoes, onions, common beans and mango. The project aimed to improve livelihoods (food security and income) and resilience through good-practice training, access to finance and quality inputs, market access and an enabling policy environment. Planned results included climate-smart practices among 30,000 households and affordable financing for 2,000 farmers. The study assessed results against baseline, relevance, the effectiveness of strategies, bottlenecks and the sustainability risks of exit.
Key findings
- Of the 327 surveyed farmers, 85.6 percent (280) sell their produce: 50.7 percent to local markets and 45.4 percent to middlemen. Other farmers were the main source of market information (49.3 percent).
- On average, 13.5 percent (44 of 327) of households worried about the adequacy or quality of food in the past four weeks. The report describes this as an improvement from 26 percent who could handle food-related shocks at project onset to 86.5 percent now not worried. Water was the basic need most often inadequately accessed (47.4 percent), followed by food (30.6 percent).
- The share of farmers growing the project's target crops fell compared with baseline: tomatoes 25.8 percent (baseline 39 percent), Irish potatoes 23.9 percent (37 percent), beans 21.2 percent (49 percent), mangoes 6.1 percent (16 percent) and onions 0.9 percent (19 percent).
- Reported cost of production per acre rose sharply from baseline: tomatoes Ksh 84,630 (from Ksh 13,142), Irish potatoes Ksh 60,369 (from Ksh 3,589), beans Ksh 13,406 (from Ksh 474), onions Ksh 93,933 (from Ksh 11,929) and mango Ksh 6,847 (from Ksh 1,780).
- According to the report, 24,875 growers had been trained by July 2023 against a target of 30,000. Among respondents, 85.3 percent recalled training on climate-smart horticultural practices, 85 percent on sustainable soil fertility management and 84.4 percent on nursery management, and 73 percent of those trained use the knowledge frequently.
- Input use was high (72.2 percent use improved seeds, 70 percent fertilisers, 69.4 percent insecticides), but only 21.4 percent do soil testing and 9.4 percent practise some form of irrigation. Adoption of soil and water conservation practices was low (for example, forest tree planting 22.6 percent and conservation agriculture 12.5 percent).
- The 33 spray service providers interviewed recalled around 90 percent of spraying good practices, and their average use of training knowledge was 91.5 percent. However, not all selected SSPs in Laikipia received the promised training, and farmers' use of SSPs remains low.
In figures
Baseline: 49 → Endline: 21.2 % of farmers
Baseline: 39 → Endline: 25.8 % of farmers
Baseline: 37 → Endline: 23.9 % of farmers
Baseline: 19 → Endline: 0.9 % of farmers
Baseline: 16 → Endline: 6.1 % of farmers
Data table
| Item | Earlier | Later |
|---|---|---|
| Beans | 49 (Baseline) | 21.2 (Endline) |
| Tomatoes | 39 (Baseline) | 25.8 (Endline) |
| Irish potatoes | 37 (Baseline) | 23.9 (Endline) |
| Onions | 19 (Baseline) | 0.9 (Endline) |
| Mangoes | 16 (Baseline) | 6.1 (Endline) |
72.2 % of farmers
70 % of farmers
69.4 % of farmers
21.4 % of farmers
9.4 % of farmers
Data table
| Item | % of farmers |
|---|---|
| Improved seeds | 72.2 |
| Fertilisers | 70 |
| Insecticides | 69.4 |
| Soil testing | 21.4 |
| Irrigation | 9.4 |
Recommendations
- Enhance access to quality inputs and financing through input financing, financial literacy and savings training, partnerships with input suppliers and affordable financing packages tailored to smallholders.
- Strengthen market linkages that connect farmers directly with buyers to reduce reliance on middlemen, and use digital platforms for timely market information.
- Invest in sustainable water management, such as rainwater harvesting and irrigation, with training on efficient water use.
- Engage local and national governments on supportive policies, such as input subsidies, favourable credit terms and better input distribution systems, and on enforcing agreed policies such as potato packaging.
- Provide continuous capacity building, including a peer mentorship system, more demonstration farms per county and training youth as spray service providers.
- Improve follow-up on centres of excellence and provide standby agronomists at each centre.
- Establish robust M&E systems with regular farmer and stakeholder feedback, and strengthen community participation and farmer associations.
Lessons learned
- Farmers in the focus groups said group training of 15-30 farmers was very effective, saving time and resources compared with training individuals.
- Field demonstrations were seen as effective because farmers learned practical skills rather than theory.
- Poor follow-up on centres of excellence weakened the sustainability of the project's interventions.
- Improved farming practices come at a cost, which the report links to the rise in production costs alongside wider economic changes.
Methodology
Mixed-methods endline using desk review, a structured household questionnaire on KoboCollect, focus group discussions and key informant interviews across the seven counties. Beneficiaries were selected by quota sampling (Krejcie and Morgan formula); 327 household interviews were achieved against a target of 330, plus 14 FGDs and 9 KIIs (the report also cites 15 KIIs in one place). Quantitative data were analysed in SPSS and qualitative data thematically.
Limitations
- Local confidentiality was not ensured, because participants were able to identify each other.
