In today’s business world, field service companies are getting busier and busier. No wonder many owners and managers are coming around to introducing more technology into their processes.
Schedule optimisation is a key area for any company looking to grow and increase productivity. After all, if you can do more jobs with the same number of engineers in fewer hours, it’s easy to see profits looming ahead.
This is why today we’re interviewing Philip Welch, routing algorithm expert and logistics data scientist, and exploring how the routing process has changed from arbitrary decisions made by an office admin to highly complex systems that take into account live traffic data and produce accurate travel time estimations. Read on to find out more about:
1. Tell us a bit more about your work on routing optimisation solutions. Where did you get your start and what made you interested in this area?
I started working at a company called MapMechanics in 2005, not long after finishing my PhD in Astrophysics. MapMechanics, now owned by CACI, specialised in route optimisation solutions. Route optimisation is a subset of an academic field called Operations Research, which is closely related to Artificial Intelligence and uses some of the same techniques. For me, I’ve always enjoyed giving computers the intelligence they need to solve hard problems, that’s mainly where my interest comes from. Also, it’s good to work in a field where you’re contributing to the fight against climate change, as more efficient route planning means less driving and less CO2 emissions.
2. You have a lot of experience in the field. How did routing optimisation and intelligent schedules look like in the beginning, when people started experimenting with the technology?
The first commercial vehicle route optimisation systems, which were released before my time, in the early 1980s, were basic but could still yield impressive results. Estimates of travel times between locations were crude as they were based on straight lines and didn’t take account of roads. Later, people started to use digital road networks to estimate travel time, and later still, live traffic updates.
Another big difference between the systems back then and now, was that the early systems created a single routing plan for the day and then followed it blindly. For many industries – particularly field service management – this is bad as you can’t serve same-day jobs or react effectively to delays due to jobs overrunning, by reassigning jobs between engineers. A much better approach is real-time route optimisation – where your routing plan continually evolves during the day as new jobs are added or delays occur. This is what our ODL Live routing engine solves - https://odllive.com/.
The key requirement to use real-time route optimisation is to automate communication between your field workers and the central planning engine – so you know where your workers are, what jobs they’ve already completed, and you can send them new jobs. The earliest route optimisation systems didn’t have that available. By the early 2000s, some companies were doing this by outfitting teams with expensive customised PDA (personal digital assistant) devices. Now smartphones are ubiquitous, so it’s a lot cheaper to develop these automated communication systems.
3. It definitely sounds like the field has grown a lot. How do people in the industries you work with feel about these innovations? Did you find them open to experimenting with new systems? Any hard resistance to change?
It’s a mixed bag. Some people are very open to trying a new system and some are resistant to change. The sad thing about route optimisation systems is because they automate the role of a human planner, adopting a system can lead to teams of human planners being significantly downsized, which obviously creates friction. Anyone who’s worried about a computer eventually replacing their job is likely to push back against the adoption of a new system.
Route optimisation systems are not usually a ‘one size fits all product’ either, they typically require some fine tuning at the start to maximise the benefits to the client’s operations. It’s very important to manage client expectations, so stakeholders understand the difference between an ‘untuned’ set of route plans they may see at the beginning of the project, and the ‘tuned’ set of plans they’ll see after suitable adjustments have been made to the routing model, based on stakeholder feedback. The hardest resistance I’ve experienced was when this wasn’t adequately explained.
4. What are 3 benefits of companies adopting this technology in your opinion? What do they stand to gain from investing in software?
- Automated control. Instead of planning every job, human planners only need to intervene in exceptional circumstances.
- Optimised routes save on mileage. This reduces both fuel costs and CO2 emissions. Companies get to save money and help save the planet.
- Increased end-customer satisfaction. With real-time optimised control you can react better when delays occur, meaning you can more reliably serve customer appointments at the time you’ve promised.
5. And finally, where do you think this technology will be in 5 years, considerings things like 5G internet speeds, the Internet of Things, and changing mindsets?
Currently, companies providing domestic repair services (e.g. electricians, gas engineers) or home deliveries, dictate appointment times to end customers and give the customer poor visibility of when they will arrive. Getting a text message saying ‘our gas engineer will arrive at your property between 7am and 6pm on Thursday’ brings a host of problems for anyone who works, has children in school, or even just needs to pop out to do the shopping. Appointment booking – and rescheduling when required – needs to be a two-way conversation between the end customer and service provider, and this is where I see the technology going in the near future.
Let’s imagine we’ve got a real-time route optimiser which can rapidly evaluate the effects of adding new jobs or changing job time slots (like Philip’s ODL Live engine or Commusoft’s vehicle tracking feature). Let’s also imagine, given the advances in AI likely over the next 5 years, that we connect a chatbot up to this real-time route optimiser. We have our customer at home, waiting for the electrician to turn up. Their child’s school rings – their child is sick and needs to be collected early. The customer sends the chatbot a text message saying they need to go out for 20 minutes, for an emergency. The chatbot interprets this message and then asks the real-time route optimiser if the job timeslot can be changed. The optimiser does the calculations and a second or two later, the chatbot texts the customer back, saying “that’s fine, we can come between 4 and 5pm instead, is that OK?”.
This may sound like magic, but the technology already exists to evaluate changes to live routing plans quickly enough for a real-time conversation with the customer. It just needs an appropriate ‘human-centred’ interface, which could be a chatbot or a much simpler UI.
All in all, route optimisation is a solution for forward-thinking businesses to monitor their assets and perfect their on-the-road behaviour, from fuel consumption to carbon emissions. Depending on your business needs as a field service company, whether it’s plumbing, heating, fire & security, etc., most can start with vehicle tracking and driving reports, which then feed into intelligent scheduling (auto-generated jobs that calculate the best route you can take to fit more bookings in a day).
For those with elaborate fleets and high-costs as well as other industries like freight carriers, on-demand deliveries, etc., a separate scheduling engine like ODL Live, that offers fast automated re-routing, simulation-based routing analysis and highly customisable scheduling rules might be the better option. At the end of the day, it’s all about the perfect business fit!