In 2018, GEOSYS’ customers have used 45% more maps than in 2017.
As we are focusing our efforts on preparing for the upcoming season, the holiday season has been the perfect time to review 2018 performance.
We strive to push the limits, year after year, of satellite-derived insight delivery for the Ag market. And today, we are pleased to share some of our key performance indicators, reflecting the trust our customers put in the data we deliver to them through our services.
GEOSYS Key performance indicators, what do they stand for?
Our customers tell us that they need recent field maps when they visit farmers or scout fields. The requirement sounds simple, but it takes a dedicated staff and a robust platform to meet their needs. At Geosys we developed four key performance indicators (KPIs) to help us measure and improve our service delivery.
Here are our KPIs:
- MAP QUALITY: This is a measure of how well the map represents the conditions of the field.
- UNIQUE DELIVERIES: This is an index based on the number of cloud-free maps delivered to customers’ fields.
- DELIVERY TIME: This is the measure of the time from satellite acquisition until the processed cloud free map is available to our customers.
- MAP CREATION TIME: This is a measure of the time it takes from registering a new field on the Geosys platform until all the current season and historical maps are ready for the customer.
To meet our customers’ expectations, we need to achieve these 4 goals simultaneously.
Let’s have a closer look at our 2018 delivery performances and at our forecasts for 2019.
Quality matters. At Geosys we use only scientific grade quality data.
It is fine to have wheel barrows full of maps but if you cannot compare one map with the next, or one field with the closest field, how do you make decisions. Or how do you automate, trigger alerts and notify Agronomists of which field to look at first?
Our customers care about the quality of the maps and analytical data they use to make decisions with their farmers.
Quality all starts with the right satellite taking a picture up-there from space. If you are into photography, you know that it is not just the gear that makes the pictures… but it is very important to have the right gear, and to use it right. In the case of satellites, the photographer and the camera are bound together:
- Sensor radiometry, Lens Geometry, signal to noise ratio, it all matters.
- Focus matters.
- Shaky hands or wobblily satellite? It matters!
- Wrong acquisition time, bad exposure…it matters.
Just like photography, where image processing can save and make a good picture from an average image, but cannot save a bad image – the same is true that you cannot make a good map if the source satellite imagery is bad. You can always try and throw Artificial Intelligence (AI) or Deep Learning at it, it will not save it. Unless you are into Abstract Art but this is not Geosys’ core business.
Even though we do not use bad quality satellites, we could still make bad maps from good imagery if we were not processing it right, just like in down-to-earth photography.
We do process the imagery that we collect with the aim of always representing reality as close as our customers need, not just for an individual map but also to be able to compare maps with each other. This takes Geometric and Radiometric corrections, Sensors inter calibration… all of it described properly in one of our earlier white papers.
This is what allows to sort through more maps automatically, thanks to change detection, and to trigger alerts and notifications.
It is not OK anymore to simply deliver truckloads of maps, we need to highlight which are the right maps to look at. We do both, through our off-the-shelf services including APIs.
We are committed to delivering more maps to provide more insights to our Customers at the right time.
We know that in-season imagery needs to be timely to be useful. The Unique Deliveries KPI counts cloud-free maps covering the entire field at a given date (one count per day even if several sensors delivered on that same day – it happens more often than you’d believe).
This past season we delivered more than 35 maps per field on average during the cropping season in the US and Canada. In South Africa, we reached an average of 48 unique maps per field. We did even better in Australia and Saudi Arabia, knowing that limited cloud-cover helped a lot over-there.
To be able to deliver these improvements, our platform processed 200% more images as compared to 2017.
Improvements in our platform and automation allowed us to increase our processing capacity.
How much did this effort yield? 3 times more imagery… 20% more cloud-free maps…
We know the extra 20% matters to our customers – it’s the difference between having a map in your hands when you need it or having nothing.
Our proprietary imagery processing system ensures our Customers can access to any available set of pixels
Providing usable data implies removing clouds and shadows that will affect the quality of the vegetation metrics on field-level maps. In cloudy areas, despite the large number of images acquired, the usable portion is fairly low and therefore the number of maps created per image is sometimes small.
Our proprietary, fully automated, process allows us to leverage any available set of pixels, even if the image is qualified by the satellite operator as very cloudy: we will capture the clear portions.
Geosys platform is sized to efficiently manage a very large volume of data
One of the satellite operators is responsible for the increase in imagery deliveries: the European Space Agency (ESA) of course, with the Sentinel 2 constellations being fully operational.
Sentinel 2 and Landsat 8 (United States Geological Survey (USGS)) represent a strong baseline of imagery sourcing that many digital Ag players are leveraging.
But the scale of our operations is significantly larger than most of our competitors. As a global player, Geosys delivers daily measures over 5 continents.
In 2018 our image library has been grown by more than 2 billion square kilometers. All of those maps are processed and available for our customers anytime.
High performing imagery service cannot rely on publicly available data sources alone. This is why we continue to improve our Virtual Constellation with addition of two new sensors to maximize coverage during key stages of crop development.
In 2018 in the US, commercial satellites (tasked) represented 36% of our map delivery.
Public data sources acquire on a fixed schedule; tasked satellites acquire on an optimized schedule, over specific areas of interest according to weather forecast to efficiently complement public data sources and deliver a significant difference. In other words, we dynamically adjust our priorities and focus depending on in-season performance delivery and business priorities allowed us to deliver more than 50% more data than public sources only.
Delivering fast to meet our Customers’ expectations
You want your maps served fresh right? In season crop monitoring requires that we deliver up-to-date insights as quickly as possible.
In 2018, we delivered access to the bulk of our in-season field maps in less than 50 hours after acquisition.
In 2019, our goal is to provide access to maps within 36 hours after acquisition, and yes, we are targeting 12 hours or less for 2021… but that is another story.
How about making maps for new fields?
How many times do our Agronomist customers have this discussion at the Farmer’s desk:
“Let’s look at this other field that we did not map yet, it’s a hybrid that’s supposed to react well to fungicide and maybe we should look at its health to make a decision.”
It’s a great discussion to have, analyzing recent maps and comparing them to crop conditions a couple of weeks back, and to field variability of past seasons… identifying which spots of the field to go look at to make the right decision.
To support the discussion, you do need the maps! All of them… and fast.
Therefore, at Geosys, we work on improving how fast we make the maps and with more maps delivered and more fields mapped every year, you get the picture.
From the moment the agronomist will finish drawing the field (or clicking on a Common Land Unit in USA) to the moment he will see the maps available to analyze, Farmer and Agronomist will not even have time to sip their coffee.
Which allows to continue working seamlessly on what matters and look at more fields and make more meaningful decisions that will impact the bottom-line.
Thank you!
All these achievements wouldn’t have been possible without the continued support of our valued Customers.
We thank them all very much and are ready to deliver more usable data points in 2019, faster and globally, to support their business growth.
Interested in easy and efficient imagery-based services?
Please visit our new API section on our website and register to get a free trial.
Your network may find this article useful, please share it on socials! We’d really appreciate it. Thank you!
