How Montreal transport skirted privacy laws

By on
How Montreal transport skirted privacy laws

Split database the key to location-based sales.

Montréal's local transit authority has launched a location-based marketing application that allows it to harvest data on its 2.5 million passengers without falling foul of Quebec's strict privacy laws.

The Société de transport de Montréal (STM) co-developed an iOS app that it hopes will help drive use of its public transport system for needs beyond the daily commute - whether that be going out for drinks after work or attending concerts and sports events.

The 'Merci' iOS app offers commuters special deals on things like coffees, tours, concert tickets and additional transport services as they move through STM's bus and metro (tube) network. It combines data the STM already knows about its passengers via their smart cards with preference and demographic data volunteered by the user in the app.

Pierre Bourbonniére, a former executive with Aeroplan — Air Canada's loyalty program — was hired by STM to find a way to introduce a loyalty program to boost use of public transport and the services it delivers to the city. 

But the project spent several years on hold while Bourbonniére negotiated the concerns of STM's lawyers and finance executives.

"It took me about four years to discover that a loyalty program was almost impossible for STM to do, because of all kinds of objections,” Bourbonniére said. 

A traditional 'points' based system was too difficult, as regulations demand "you need to have in the bank the dollar value of the points you give to customers, because in the eyes of the regulator, today you have a debt towards them and one day you have to pay that debt," he said. "In the public sector, this is not allowed."

"Further, Quebec is one of the most strict provinces in North America in terms of privacy," he said. "These two things alone kept the project on the shelf for so long."

It was a fresh round of discussions between SAP, STM technologists, accountants and lawyers that finally resulted in a breakthrough.

"We've come up with a bit of a miracle," Bourbonniére said. "The key element we came up with was the idea of an awards program with instant gratification. We want to gratify customers as soon as they get into the [transport] system.

A legal and technology solution

So how did the technologists at SAP and STM convince the agency's lawyers to sign off on the solution?

They proposed a backend for the app using ring-fenced data sets that are brought together for only a split second at a time.

The back-end of the application splits traveller data into two separate databases, with key personal details 'de-identified' to the point that access to any one of the two databases could not identify an individual.

STM already knows some information about commuters courtesy of an integrated smart card ticketing system called Opus that was introduced in 2008.

This information resides in a legacy CRM database in Montréal which contains only a customer's name, ID number, their Opus card details (fares purchased, public transport habits) and the first few digits of their postcode. The 'habits' include where commuters enter the system, but not necessarily where they exit.

When commuters download the 'Merci' iOS app, it separately asks them for more personal details - their age and sex, their consumer habits and preferences. Rather than store this in the same database as the Opus card details, this preference data is stored offsite in SAP's Waldorf (German) data centre.

"All the information about a customer's preferences and buying habits cannot be in the same database as the mission-critical information about the customer," said Bourbonniére. 

"The lawyers would forbid that. There has to be a Chinese wall."

The Merci app uses data from both sources. The two customer records only come together for a split-second at the point at which a user of the Merci app selects an offer.

"The moment a customer chooses an offer - imagine an atom smasher - this information comes together within a fraction of a second."

STM has signed up a network of 340 partners - a mix of entertainment providers (sports stadiums, theatres, tours etc), food and beverage outlets and retailers - that use the application to offer special discounts and one-off treats to commuters as they move in and out of the transport system. Other transport providers - such as the city's bike rental system and privately-owned regional bus services like Keolis - also offer discounts and specials via the Merci app.

The marketers (STM's partners) don't know the particulars (name) of this individual that took on the offer, but do know the defining characteristics of the individual.

"We get the one-to-one goal of delivering value to our customer, but the preference data alone can't identify the individual," Bourbonniére said.

Among the most clever uses of the application: 

  • A book store that offers commuters moving toward its store the opportunity to download a sample chapter of a book that is on sale, in the hope they'll get hooked and buy it.
  • The Montreal Opera House which offers special prices on unsold (and thus perishable) seats at 4pm the afternoon of the performance.
  • Supermarkets that offer their specials on bus routes for commuters soon to pass by who are likely to be thinking about what to cook for dinner.
  • Food trucks can let patrons know when they are parked close to their office, should they be up for a lunchtime treat.

Learned behaviours

On first use of the Merci app, it doesn't behave much differently to existing location-based marketing applications (such as Closebuys, ShopperNova, Spreets), and it relies heavily on native iOS features (maps API, location-awareness, Passport etc).

It's only after regular use, says SAP's Herve Pluche, VP of Precision Marketing for SAP, that customers might notice the key difference. 

SAP's portion of the app backend is based on SAP's HANA in-memory database and the German software company claims it is capable of 'learning' over time what offers a commuter or profile of commuters use most often, and pushes those to the top of the list.

"It is self-learning - the more you interact with us, the more history we have, the more we understand your interests as measured by actions, the more likely we are to give you information you care about," Pluche said.

Precision Marketing is the brand SAP will use to on-sell what it is co-developing with STM to other customers. SAP retains the intellectual property from the project and is free to re-sell the same solution to other transport companies or other organisations with an ecosystem of large numbers of customers and partners.

Pluche said the siloed structure of the data backend is a "breakthrough in terms of compliance to privacy regulations and meeting marketing strategies".

"The idea of breaking down information into separate silos that can be combined only when a customer initiates a session means there is no latency of information once a session is complete," Pluche said.

Read on for the results and the road ahead...

The STM Merci application was released to 20,000 beta testers in May 2013 and around 1000 offers have been made available to these early adopters. It will now be promoted to all 2.5 million users in the hope that 100,000 residents will be using it by 2014.

STM introduced push notifications within the Merci app one month ago and is starting to calculate the results. 

The transport agency started conservatively - it will not send more than one notification to any given customer within 36 hours.

Of the 20,000 beta testers, 17,900 have kept using the app within the first week - which is far higher than average. 

One in two users gave STM "the full monty" — filling out all preference data to drive better deals. Around one in three refused to give any preference information at all. 

Customers are logging in at an average of six sessions a week.

The average open rate is around four percent — and the conversion rate of that four percent is 27 percent. The top 25 percent of offers attracted open rates of over 40 percent.

"Achieving double-digit conversion for the best offers is an excellent result - but the spread suggests there is room for learning,” Pluche said. “The learning curve for marketers is extremely steep. We didn't know which offer would work best - understanding the why is helping CMOs [chief marketing officers] write the book about real-time interaction with consumers at the point of decision.”

Under the co-development deal, STM pays SAP a license fee based on 'results' - the number of active users and the actions they take (redeeming an offer or marking an offer as a 'favourite', for example). STM in turn gets additional revenue from the Quebec Government the more 'rides' it calculates on its system.

At present, retail partners can make offers on 'Merci' free of charge. But this is likely to change if the project hits scale. "We are looking at recouping the investment by asking partners to pay to be in the program on a click-through basis,” Bourbonniére said.

The 'Merci' app at this stage remains available only on iOS, and the STM expects it won't be until mid-late 2014 that it's available on Android devices.

Further, the app requires strong 3G or Wi-Fi capability and the STM Metro system currently offers none when trams move underground. Local carriers have committed to rolling out 3G in these areas.

Finally, the app has never been 'audited' by any authority - only cleared by the STM's lawyers, who assumedly have sound knowledge of Quebec privacy legislation.

"What we've learned is that there is a myth around customer privacy concerns," Pluche said. 

"Paranoia about privacy gets lifted when you do two things — when you have trust in the brand, which riders have in STM; and when you have relevance, in which a customer is given value instantly in return for sharing their information."

Multi page
Got a news tip for our journalists? Share it with us anonymously here.
Brett Winterford

One of Australia’s most experienced technology journalists, former iTnews Group Editor Brett Winterford has written about the business of technology for 15 years.

Awarded Business Journalist and Technology Journalist of the year at the 2004 ITjourno awards and Editor of the Year at the 2009 Publishers Australia 'Bell' awards, Winterford has extensive experience in both the business and technology press, writing for such publications as the Australian Financial Review and The Sydney Morning Herald.

As editor of iTnews Brett has led a team of award-winning journalists; delivered speeches at industry events; authored, commissioned and edited research papers, curated technology conferences [The iTnews Executive Summit and Australian Data Centre Strategy Summit and also shares the judging of the annual Benchmark Awards.

Brett's areas of specialty include enterprise software, cloud computing and IT services.

Read more from this blog: System II

Most Read Articles

Log In

Username / Email:
  |  Forgot your password?