Monday, February 5, 2024

SIMPLYGO FIASCO : THE BIG DATA AMBITIONS AND $40M DILEMMMA EXPLAINED


"Public sentiment is everything. With it, nothing can fail. Without it, nothing can succeed." ... Abraham Lincoln

Land Transport Authority of Singapore had eggs in their face in the aborted migration of card-based transit ticketing system (CBT) to account-based transit ticketing (ABT). CBT cards - Ezlink and NetsFlashPay, were to be replaced by ABT card - SimplyGo-Ezlink (SimplyGo for short) by1 June 2024. Barely two weeks after announcing the dateline, government walked back its planned move in response to public outcry. Transport Minister Chee Hong Tat went so far as to say the planned move was a "judgement error".

The criticisms were centred on two main issues. The new SimplyGo card does not display fare charge and credit balance on the train and bus card readers and it cannot be used for ERP (electronic road pricing) and carparking.

The public condemns SimplyGo as a retrogression, a backward move. In fact, a retro word doesn't quite capture the essence of the fiasco. The Germans have a better word -- verschlimmbesserung. This means making a change for the better but end up worse than before.

However, people are seeing the card for the system. The SimplyGo system is the whole backend infrastructure, developed and managed by Acclivis Technologies, that enables payment by credit cards, bank cards, mobile and wearables, SimplyGo card, as well as cloud storage of commuters' accounts, transaction data, and big data analytics.

It has been one huge PR disaster for the government and explanations so far have not been crystal clear. Chee's "end of life" speech at the press conference was obfuscating. He explained the June dateline to migrate to SimplyGo was due to the ROD (run out date) of the CBT system. A further S$40m for hardware and maintenance is required to allow the CBT system to continue till 2030.

Folks are incensed additional $40m tax money to be spent is system design error but Chee seemed to place blame on the public. Because you complained, more spending is needed. The public has got it wrong. It does not cost anything to retain the use of EZlink and NETsFlashPay cards. What Chee meant was after a series of focus group discussions from 2020 to 2023 with about 1,000 commuters from different groups, they realised older generation folks have difficulties with the use of apps in SimplyGo. For this reason it was decided to retain the use of CBT Concession cards till 2030 and this requires additional spending of S$40m. The government felt this was necessary spending. So the S$40m additional cost is independent of the decision on the fate of EZlink and NETsFlashPay cards. Whether these cards are discontinued or retained, the $40m will sill have to be spent.

The same card readers accept both CBT and ABT cards as well as mobile and wearable devices, so most folks are perplexed about the S$40m spending. What Chee did not explain is the topping up of stored values. SimplyGo cards are topped up via the app or Web. EZlink, NETsFlashPay, Concession cards require top up kiosks. In order to continue using CBT cards, S$40m is needed for additional kiosks for the new MRT stations being built, replacement of hundreds of existing aged ones reaching end of economic life, and maintenance.

Why are top up kiosks required for CBT cards? This is because stored values are encoded into the microchips by these kiosks. The debit and credit computation of CBT cards are done in the microchips which is why card balances can be displayed on readers immediately. Stored value or balance is credited or increased when commuter tops up, and decreased or debited when commuter taps out.

On the issue of non-display of fare charge and card balances, the revelation of focus group discussions in 2020-2023 is a shocker to many. A product review is normal. However, SimplyGo went live in 2019. Fundamental weaknesses should have been uncovered at much earlier product specification stage.

Minister Chee explained SimplyGo does its computational work in the backend system and is thus unable to display fare charge and card balance at card readers. It has the info but to download it to the card readers take some delays not acceptable in a fast moving commuter traffic situation.

Let's consider how the system works. When a commuter taps in, the card reader sends certain info via internet to the backend for authorisation. If it's a credit card or bank card, a message goes to the issuing bank via the processor's network, to confirm availability of credit. The issuing bank responds accordingly. If it's a SimplyGo card, the system checks for card balance in commuter's account. Whether approved or denied, the backend sends the info to the readers. In the case of trains, the turnstile is released for commuter to enter. If authourised, the system keeps the transaction open. When commuter taps out, certain info is sent over to the backend to allow the transaction to be closed. In the case of trains, the turnstile is released for commuter to exit.

These messaging here and there takes place almost instantly. So the question is, when a commuter taps in, how long will it take the backend to check a SimplyGo card balance and send the info to display on the reader. And on tapping out, how long will it take for the backend to update and send the info to the reader so he knows the fare and balance upon exit. Is it nanosecond, microsecond, or millisecond?

I can offer an explanation here but note it is speculative. SimplyGo is probably not updating accounts in real time. That is why Minister Chee is technically correct that SimplyGo can show balances on card displays but it takes too long. This also explains one of the complaints of many commuters that the system is slow in updating their accounts. 

In evaluating systems, outside of cost issues, we see features in two groups - 'must have' and 'nice to have'. For commuters, the SimplyGo card must have fare and card balances displayed on card readers immediately. All other features the government pointed out such as cancel lost cards, top up stored values on the go, topping up for someone else, checking travel expenses, dynamic pricing etc,  are just nice to have.

For the government, it sets grand plans. In this instance it is all about 'smart city'. In the Smart Cities Index Org report for 2022, Singapore was ranked 26th. London holds top spot. Basically, smart city is about the application of Information Communication Technology in urban development. The government has road maps for various aspects. In transit systems, the envelope to punch is MaaS (Mobility as a System) which is one single platform for trains, buses, taxis, ride-hailing services, electronic road pricing, carparking and bike-hailing services. The road to this is via SaaS (Software as a Service) which is to use a cloud service provider, taking away the IT function from transit operators. This approach seeks to abandon monolithic closed loop proprietary systems for a fee-based system from a specialist service provider and allow transit operators to focus solely on running the trains and busses instead of ticketing technology. Acclivis Technologies, a Singapore SaaS wholly owned by Hongkong Telecom, designed, built, and manages the SimplyGo system.

In his 27 Jan 2024 Gutzy article, Terry XU said SimplyGo is a classical case of technology for technology sake. Is this fair comment? What does the government want out of a transit ticketing system? In two words - Big Data. Transaction data captured by the system provides rich information on transportation patterns which serves urban planning and development. CBT systems do not capture transaction data. 

In the case of credit and bank cards, transaction data is captured by SimplyGo. For stored-value cards like EZlink, NETsFlashPay, and concession cards, transaction data is not captured. That is the underlying reason why CBT cards have to be phased out.

It is not a problem to have SimplyGo do transaction data capture for CBT cards in same manner as for credit and bank cards. The problem is CBT cards are not personalised so the data has limited utility absence demographics and locations details. The simple solution is personalise EZlink and NETsFlashPay cards similar to credit and bank cards. However, this solution requires card value top up kiosks to be retained which carries a high cost (old ones need replacement, additional ones for new stations, maintenance).

ABT or account-based ticketing provides the solution to doing away with the top up kiosks. Enter SimplyGo cards. Personalised. Transaction data capture. Top up on the go without kiosks.

The only problem is SimplyGo cards do not display fare charge and card value balance on readers. As mentioned, transactions of SimplyGo cards are not updated in real-time. Will real-time update solve the problem? Perhaps.

Another issue the public is concerned is the cost of SimplyGo system which LTA is not sharing. An SaaS project would mean LTA bears no development cost and the vendor recovers cost in a fee structure that is generally higher in the first few years and subsequently lowered thereafter. By accepting credit and bank cards, transit operators are merchants with several charges to take care off, such as processors' network charge, ACH (automated clearing house) charges, etc, on top of SaaS fees. Transit fares are super micro payments in very high volumes. Fortunately, the various card processors allow for aggregation which reduces volume and cost. In fact, without aggregation, it is economically not feasible to use credit and bank cards for transit ticketing. How much of these increased ticketing cost has been worked into fare increases is the question.

For more than 50 years, the PAP has normalised a nanny state governance where, perhaps well-intentioned, tough regulations have been rammed down on the public. There is cocky confidence the public will bite the bullet and negative sentiments will be numbed with passage of time. This SimplyGo farce has been a plebiscite on the need for immediate display of fare charge and stored value balance on fare card readers. The admission of 'judgement error' is a big deal for Singaporeans. It is a paradigm shift in the government's policy execution that public voice matters. It is good this time the powers that be had chosen to heed Abraham Lincoln.


Note:
I am currently parallel publishing at Substack. Do visit and subscribe here

This platform has withdrawn it's subscriber widget. If you like blogs like this and wish to know whenever there is a new post, click the button to my FB and follow me there. I usually intro my new blogs there. Thanks.