Welcome to the first post on the new goto blog. Over the next few months, our team will be sharing ideas, opening up data and providing insight on how we design and implement new features.

In this post we want to take a look at why we believe small jurisdictions (like Jersey) need a taxi platform, not a taxi app, and share the development process behind goto. Before trying to answer this, it's important to define what we mean by platform and app.

Taxi app: an app granting access to drivers associated with a single company or organisation
Taxi platform: an app granting access to drivers associated with multiple independent companies

Whether an app or a platform, both are examples of a two-sided marketplace, where the service facilitates economic exchange between two distinct user groups that provide each other with the benefits of a large network. In this case, the user groups are "taxi drivers" and "people looking for a taxi", and the benefit provided for each group is easy access to the other. To successfully build and maintain a two-sided marketplace it's important to match demand and supply, on both sides. If users' requests aren't answered, they will stop using your app, and if drivers can't get enough work, they'll look elsewhere.

In small jurisdictions (like Jersey) no single taxi company can service the entire island. Anyone who has tried to get home after a busy night out can testify to this, with it usually taking three or four calls to find a company with vehicles available. As industries around the world are called to modernise, many taxi companies have started launching their own apps, allowing users to request a ride without picking up the phone.

While it’s fairly easy to dial a new number and try again, downloading an app for each company is a whole different ball game. For each app, users have to sign up, add their card details, fill out a request, check between them, then finally cancel any open requests when they get a ride. Each app downloaded is another app asking for payment details, personal information, and location data. Need to get a receipt from your last ride? Unless you remember which app you used, that means searching through your trip history in 4+ different apps. Got a new debit card? Not only is that 4+ apps to update, but 4x the risk of a security breach potentially impacting financial and location data.

We couldn't do a blog post from Jersey without some examples, so let's take a look at another local app, the JERSEY TAXIAPP, operated by the Jersey Taxi Drivers Association (JTDA).  The JERSEY TAXIAPP was launched in November, and since they launched we've been automatically collecting data on driver availability, wait times and vehicles.

JERSEY TAXIAPP / Availability Heatmap / © goto.

In this chart, each vertical column is a day, beginning at 00:00 (top) and ending at 23:59 (bottom), and each row is one hour. Areas in red indicate periods where no drivers were available. As you can see from this chart, the JERSEY TAXIAPP is not available for significant portions of the day, and often has zero availability in the early hours of the morning (when demand is highest).

Let’s take a look at another local company, Liberty Cabs. Here is the chart for Liberty Cabs, for the same period:

LIBERTY CABS / Availability Heatmap / © goto.

Liberty Cabs is one of the largest private hire companies on the island, but even with their considerable resources there are still large patches of intermittent service. As Liberty Cabs display their vehicles on the map, we can also chart the number of vehicles they have available:

LIBERTY CABS / Available Vehicles / © goto.

The JTDA have disabled the feature showing the number of vehicles online so we don’t have this data for the JERSEY TAXIAPP, however we can infer that the JERSEY TAXIAPP likely has significantly less than 15 active drivers, as Liberty Cabs has a far higher availability with an average daily peak of 15 drivers. Our analysis puts the active driver count for the JERSEY TAXIAPP in the range of 6-10.

With the other two local apps having around 30 drivers combined, it’s not surprising that neither have taken off. In fact, it would be almost impossible for any local taxi company to launch and grow an app successfully – no single company has enough drivers. This is where the benefits of a platform solution become evident – you’re not limited to the drivers for a single company.

Due to the complexities involved, a platform solution is significantly harder to launch than an app. For one, a platform solution has many more stakeholders, and it's almost impossible to find an off-the-shelf solution. In a place like Jersey, you have four major stakeholders: regulators, users, drivers and dispatchers (or booking entities).

Before we started work we met with regulators from all branches of government to evaluate their needs. Across every department, the most common request was access to anonymized data for unmet demand and traffic flow surveys. At the moment, regulators collect data on the industry through secret shopper surveys, which are incredibly expensive and labor intensive. Access to live data on wait times, vehicle availability and demand hotspots would be game changing, as none of this data exists currently.

We went on to identify four key user demands: in-app payments, push notifications, driver ratings and live support. With drivers, the demands were similar: easy payments, report passenger issues and live support. There were also a couple we didn’t expect, like night mode, which inverts light colors to make prolonged viewing in the dark easier on your eyes.

Finally, we organised meetings with all but one of the major local dispatchers. We fielded a wide range of requests, but the core business requirements were simple: support their current tariffs, protect their brand image, honor their existing fare split agreements and provide online dispatching tools. It was at this point (early Feb ‘18) that we reached out to the JTDA for the first time, hoping to get their input and share what we had been working on. Unfortunately they refused to meet us, and have ignored all our attempts to reach out since then.

After reviewing the business requirements, and speaking with several vendors, we reached the conclusion that there were no off-the-shelf solutions which fit all the requirements, and started putting together a team to build it from scratch. The first line of code was written towards the end of Feb ‘18. Throughout this process we met with dispatchers, drivers and users to demo the app, and seek feedback. By October, we had our first beta version. Working with the dispatchers that helped us design it, we started doing small-scale tests, inviting family and friends to be the first passengers. Results were incredibly successful, with great feedback from passengers and drivers. We made a few more changes, and officially launched in early December.

Unlike single-company solutions like TaxiCaller, goto has no monthly fee – it’s completely free to sign up and use (we take a 5% cut of completed rides). When drivers go on holiday, they don’t pay £10/week for something they’re not using. Dispatchers can brand their drivers app, and users see their colors and logo when in a ride. On top of that, in-app payments are included automatically, with all fees included in our fee.

We hope this data and insight into our development process has been helpful. If this blog post has inspired you to start driving using goto, you can find our more about what we offer drivers. If you’d like to get a ride using goto, you can download the rider app by searching ‘goto rider’ or clicking here.

As always, we welcome any discussion, regardless of your view (that said, please keep comments civil). You can comment on this post via Facebook at this link, or you can contact us directly with your comments via the chat in the bottom right, or via Facebook Messenger.

We're also happy to share the data behind the charts in this post. If you'd like access to this data, please get in touch with the team.