leaf_logo.png

Better Prices for Your Produce: Hard Numbers for the Value of AI-Automated Harvest Yield Forecasts

March 5, 2019 2:40:16 PM EST
 

money-2696219_1280

Each greenhouse is unique. Your cultivation methods, equipment, and size of workforce and facilities are but a few factors that make you distinct from your nearest neighbouring greenhouse. Given the unique attributes of your business, it is challenging to calculate the exact value and return on investment (ROI) your greenhouse will obtain from Motorleaf’s AI-automated harvest forecasting services unless you provide us with detailed information.

You can make a detailed, confidential value gain/ROI calculation yourself at a later time. Until then, use this estimate as a guide. Here we focus on the value gained from optimal produce prices by our AI automation technology derived from information and observations from a 3-hectare greenhouse producing tomatoes.

At the end of this article you will find detailed spreadsheets and information about the greenhouse and it's production methods (see the full spreadsheet here).

 

More accurate harvest forecasts provides optimal produce pricing


Cost savings per hectare per season: € 7 493.40

Untitled drawing

Predicting weekly harvest yields proves challenging, where this European grower averages a prediction error of almost 19% over or under the actual weekly yield. When the greenhouse under produces to meet contract obligations, they buy the missing produce at an average of a 15% price-increase premium over the weekly market price. In the case of overproduction, the greenhouse tries to sell the perishable produce quickly, selling the excess at a price of 50% less than the weekly market price.   

Under-projecting and the resultant sale of excess tomatoes at a discount costs the greenhouse about €38 752 per harvest cycle. Overprojecting and the resultant purchase of missing tomatoes at a premium costs the greenhouse about €3 515 per cycle.

Precise harvest forecasts by Motorleaf reduce error in estimating future yields by half, which in this case is about 9.3% average error from the actual weekly yield. Such precision would enable the greenhouse to know early on how much excess or deficient produce they will have, meaning the greenhouse is better positioned to buy and sell their over- and under-produced tomatoes at market price; the greenhouse also gains a more confident calculation of the amount of excess or deficient tomatoes they must sell or buy, which helps avoid unnecessary produce purchases at suboptimal prices.

If the greenhouse used Motorleaf’s harvest forecasts, lowered under-projecting and the resultant sale of excess tomatoes at a discount costs the greenhouse about €16 303 per harvest cycle; less over-projection and the resultant purchase of missing tomatoes at a premium now would cost the greenhouse about €3 484 per cycle. The total cost savings from Motorleaf’s precise harvest forecasts are thus €22 480 for this harvest cycle.

Total cost savings for the greenhouse is thus € 7 493.40 per hectare per season; this is roughly € 1 500 in cost savings per hectare per month during harvesting. These cost savings do no include the value gained from automating harvest yield forecasting, which enables your team of growers to focus on other value-added work such as optimizing grow protocols or testing new varieties (stay tuned for a future blog article that calculates the value from this automation).
 

About the greenhouse and their produce sales

A 3-ha European grower of tomatoes provided the sales data and weekly produce volumes used to make the calculations herein, which you can view in detail below; the values represent one harvest cycle without intercropping. We anonymized the information and multiplied the produce prices and yields by a common factor in order to protect the identity of the greenhouse facility.

Tomato yield and prices here are for a common and affordable tomato on the vine (TOV) variety. Expect to obtain a higher value gain and return on investment when implementing our technology for tomato varieties sold at higher market prices.

Our customers obtain on average a 50% reduction in error in forecasting weekly harvests within the first year of implementing our AI technology; we use this average error reduction here. Note that our technology becomes more precise when trained with additional greenhouse data from each harvest cycle, where our customers can obtain >70% reduction in error for harvest forecasts after one year. Thus, expect to obtain even greater value gains from our technology after one year.

(Follow this link to view the following spreadsheets in 'landscape format', meaning, as a full page.)

Get a free consultation about our AI automation services today by filling out this short form.

Submit Form 

Topics: greenhouse, tomato, automation, vegetable, yield prediction

Subscribe to Email Updates