The Problem & The Brief

Sheep farming is a business of thin margins and hard variables. Weather, disease, predation, theft, any one of them can turn a profitable season into a loss. And yet, for most farmers, the tools available to manage these risks hadn't changed much in decades. Livestock monitoring meant walking fields, counting heads, and hoping nothing had gone wrong overnight.

Spaze came to Hatton Locks with a clear ambition: build a GPS-based livestock tracking system that farmers could afford, that would work reliably in rural environments, and that would deliver a return on investment measurable in real money saved.

The brief wasn't just to build hardware. It was to bring an entire product to market, from the collar on the sheep to the invoice in the farmer's inbox. That meant designing the tracking technology, building the network infrastructure to support it, and wrapping it all in a commercial platform that Spaze could operate independently once Hatton Locks stepped back.

The guiding principle throughout the engagement was straightforward: needs must first, wants second. Spaze needed a working product, a functioning business, and the capability to run it without us. Everything else was secondary. That discipline kept the project focused, kept costs under control, and meant that when the handover came, it was clean.

What made this project genuinely interesting, and genuinely difficult, was the environment. LoRa WAN networks in rural Wales don't behave the way they do in a city. Sheep don't stay still, don't stay in range, and have a remarkable talent for destroying anything attached to them. And farmers, rightly, have little patience for technology that promises more than it delivers.

The system had to work in the real world, not just on a spec sheet.

The core goals were simple to state, harder to execute. Reduce sheep losses from illness, theft, and predation. Give farmers visibility into flock movement and grazing behaviour. Lower veterinary and operational costs through smarter, data-driven decision-making. And do all of it at a price point that made commercial sense for a small farming business.

What emerged over the course of the project was something more valuable than any of us had initially anticipated. The tracking system worked. But the bigger discovery was what farmers did with the data once they had it, and how quickly it became clear that the real value wasn't in finding lost sheep. It was in stopping them from getting sick in the first place.

That story starts with the hardware.

​

Building the Hardware

The collar is where the product lives or dies. Everything else in the system, the network, the platform, the data, depends on a piece of hardware that can survive on a sheep's neck for over a year, in all weathers, without being removed, damaged, or running out of power.

Getting that right required making a series of deliberate trade-offs, and Odoo's supplier and inventory management tools played a more important role in that process than you might expect.

The core of each collar was built around the Heltec CubeCell development board. This wasn't the smallest option available, the Onethinx module would have been more compact and marginally more power-efficient, but the Onethinx came with a significantly higher unit cost and no established supplier relationship. At scale, across hundreds of collars, that cost difference matters enormously. The CubeCell offered integrated LoRaWAN connectivity, solid power management, and a price point that made the product commercially viable. It was the right call.

For GPS, the Quectel L76-L was selected over the more widely used Ublox NEO-6M and NEO-7M modules. The L76-L's advantage was twofold: lower power consumption and faster signal acquisition. In a system designed around brief, scheduled GPS fixes rather than continuous tracking, the ability to get a satellite lock quickly and then power down again was critical to battery life. The Ublox modules, while readily available and well-documented, drew more power and were slower to acquire a fix, a meaningful disadvantage when every milliamp counts.

The collar itself was built around heavy-duty nylon webbing with a moulded ABS compartment housing the electronics, stitched in rather than clipped. Tamper-proof Torx-style bolts secured the assembly, making removal without the correct tool deliberately difficult. Sheep will try to remove anything unfamiliar. The collar design accounted for that, balancing security with the animal welfare requirement that nothing caused discomfort or restricted movement.

Establishing the supply chain for all of this was handled through Odoo from day one. Supplier contacts, component pricing, order history, and inventory levels were all tracked centrally, which mattered because none of these supplier relationships existed at the start of the project. Building them from scratch while also managing development timelines required a level of procurement discipline that would have been difficult to maintain without a proper system behind it.

The result was a collar that could last over a year on a single charge, withstand the physical demands of farm life, and be deployed at a cost that made sense for the farmers buying it.

None of that happened by accident. It happened because every component decision was evaluated against three criteria: does it work reliably, can we source it consistently, and does it keep the unit cost where it needs to be.

The Network & Power Strategy

Building a GPS tracking collar that lasts a year on a single battery sounds impressive. Making it true requires solving a problem that has nothing to do with the collar itself, it requires solving the network.

The system used LoRa WAN, a low-power wide-area network protocol well-suited to rural environments where devices need to transmit small amounts of data over long distances without consuming significant power. The gateways chosen were the Heltec MT-01, deployed in weatherproof Altelix enclosures alongside a Raspberry Pi Zero W for local management. Mast-mounted with non-directional antennas, a single gateway could cover a radius of up to 15km in flat terrain, meaning most farms needed only one.

That coverage radius was not guesswork. Each installation used VE2DBE radio propagation modelling to calculate link budgets and generate site-specific coverage maps before a single gateway was installed. Hilly terrain reduced line-of-sight range, but elevated placement compensated. The point was that every installation was designed, not assumed.

Operating the network legally required an Ofcom licence. The licensing model used was priced per installation rather than as a revenue-based variable cost, which meant farmers faced predictable, fixed regulatory expenses, an important detail when selling to businesses that plan on tight margins.

The power strategy was where the real engineering discipline showed. GPS receivers and LoRaWAN transmitters are both high-drain components. Running either continuously would flatten a collar battery within weeks. The solution was aggressive scheduling: the GPS module woke at defined intervals, acquired a fix, and powered down. Data was stored locally and transmitted once per day in a single burst, rather than sent with each fix.

The trade-off was latency. Configuration changes took up to 24 hours to propagate, because collars only received new instructions during their daily upload window. In an agricultural context, where nothing happens on a minute-by-minute basis, this was entirely acceptable. Farmers planned ahead anyway.

The Heltec CubeCell's 8-channel gateway architecture introduced a scheduling consideration: only 8 devices could communicate simultaneously. With average Welsh flock sizes around 650 animals, transmission timeslots had to be distributed across the 24-hour window. SiteWhere, the open-source IoT platform handling data ingestion and device management, coordinated this scheduling. Collar swap-outs and new device registrations were project-managed through Odoo, ensuring new units were slotted into the transmission schedule without disrupting existing devices.

The combined effect of optimised gateway placement, scheduled transmissions, and deep sleep cycles meant that extending battery life required no increase in battery size. A larger battery would have increased unit cost and collar weight. Better network design achieved the same result for a fraction of the cost.

This is the kind of decision that doesn't appear on a product brochure but makes a significant difference to the economics of running the business, and to the farmers paying for it.

Turning Data into Farm Value

The original pitch for a livestock tracking system was straightforward: know where your sheep are, lose fewer of them. It's a reasonable value proposition. It turned out to be the least interesting one the system could make.

Once farmers had access to GPS movement data, the most significant financial benefit wasn't loss prevention. It was disease control.

Sheep vaccinations range from £5 to £54 per dose. For a flock of several hundred animals, blanket vaccination programmes represent a substantial recurring cost, and historically, farmers had little choice but to vaccinate broadly, because there was no way to know precisely which animals had been in contact with an infected one. GPS-based contact tracing changed that. By analysing movement histories, farmers could identify exactly which sheep had spent time near a sick animal, vaccinate only those animals, and leave the rest untreated. The cost savings were significant and immediate.

One farmer using the system noticed that sheep grazing in a particular field were falling ill at a noticeably higher rate than the rest of the flock. Water contamination was ruled out. Analysis of movement data and health records pointed clearly to a specific grazing area. Physical inspection of that field found ragwort, a plant toxic to livestock that causes progressive liver damage. The field was fenced off, the illness pattern stopped, and the farmer had a data trail that would support any insurance or veterinary inquiry.

Without GPS data, that field might have caused losses for years before the connection was made.

Grazing optimisation emerged as another application. By overlaying movement data with pasture health assessments, farmers could identify overgrazed areas, plan rotation schedules more effectively, and extend the productive life of their land. Activity patterns also proved useful as indirect shearing records, sheep move differently after shearing, and farmers began using the change in movement signatures to track when individual animals had been through the process.

None of these applications were specified in the original brief. They emerged from farmers having access to data and finding their own ways to use it.

SiteWhere handled the processing of raw GPS data, filtering anomalies, applying rules, and structuring outputs. Odoo delivered the results to farmers through a customer portal that gave them daily movement reports, automated alerts for stationary animals (a potential indicator of illness or injury), and access to their full location history. The alerts were the feature farmers came to rely on most: rather than logging in to check a dashboard, they received a notification when something needed their attention.

The commercial lesson here is important. The product that sold was GPS tracking. The product that retained customers and generated referrals was a farm management tool that happened to use GPS. Understanding that distinction changed how Spaze marketed and priced the system, and it came directly from watching how farmers used what had been built for them.

The Business Behind the Product

A GPS collar that works in the field is a hardware project. A GPS livestock tracking business that generates revenue, supports customers, and grows without constant intervention from its founders, that's a different problem entirely, and it's the one Hatton Locks was ultimately there to solve.

Odoo was chosen as the operational backbone because it could handle the full commercial lifecycle of the product in a single platform. That mattered for Spaze, which was a small team with limited resource. A business running on disconnected tools, one system for invoicing, another for customer management, a spreadsheet for inventory, would have created administrative overhead that a small team couldn't sustain.

Sales and customer onboarding were managed through Odoo's CRM and sales modules. Farmers could order equipment, modify service contracts, and manage their accounts through a self-service portal, reducing the support burden on Spaze's team and giving customers the autonomy they preferred. Automated invoicing handled renewals without manual intervention. Inventory management tracked collar stock, gateway hardware, and component levels, feeding directly into procurement decisions.

The self-provisioning capability was particularly valuable. Spaze was small. Putting routine account management, hardware ordering, and service modification directly in the hands of farmers meant the team could focus on growth rather than administration.

Beyond operations, Odoo stored all product documentation, price books, document templates, and training materials. A customer training module was built directly into the platform, covering system use for both farmers and Spaze's own staff. Internal documentation covered sales processes, troubleshooting procedures, and product specifications. When Hatton Locks concluded its engagement, Spaze didn't just have a product, they had a fully documented, operationally functional business.

That was the point. The handover was the deliverable, not an afterthought.

Open-source tooling played an important role in keeping development costs down. SiteWhere, the IoT data platform, was open source. The LoRaWAN stack was built on open standards. Using existing, proven platforms meant Spaze didn't carry unnecessary licensing costs in the early stages of the business and retained the flexibility to migrate or rebuild components as the business scaled, with full understanding of how the system worked, rather than dependency on a black box.

Spaze entered the market with a product that worked, a business platform that could scale, and documentation that meant the team didn't need to rediscover decisions that had already been made. Competitors who followed an MVP approach, getting to market quickly with fewer capabilities, consistently had to reinvest to catch up with features the Spaze system had from day one.

That's the case for doing it properly from the start. Not over-engineering, not gold-plating, but making sure the foundations are solid before the business is built on top of them.

If you're working on an IoT product, a precision agriculture solution, or any technology that needs a commercial platform built around it, that's exactly the kind of problem Hatton Locks exists to solve.