Pages

Pages

Friday, September 14, 2012

How to save NFC: Kill the idea of mobile payments & operator involvement

There's been lots of hand-wringing over Apple's decision to exclude NFC support from the iPhone 5. It's not because it can't. It's because it won't. Apple's marketing VP Phil Schiller is quoted as saying It’s not clear that NFC is the solution to any current problem”.

Spot on.

I've been critical of NFC for some considerable time, and it feels pleasing to be vindicated by Apple's doubtless consumer-centric and design-first approach.

I see three main problems with NFC:

1) Focus on mobile payments & other transaction-based use cases
2) Complexities around the secure element stemming from telcos' insistence on being involved in the NFC value chain
3) Ergonomic deficiencies.

The third one is easiest to explain. Simply: tapping a piece of expensive, glass-encased electronics on solid objects is stupid. Furthermore, making people interrupt whatever they're doing on a phone to buy something / get on a train / whatever is equally stupid. We all multitask. Let me use my phone & a card/cash simultaneously.

1) and 2) are linked. The belief that the "killer app" for NFC is paying for things - or other "monetised" apps like ticketing - has led mobile operators to say "we want a slice of that!". This has then led to interminable wrangling over the architecture for security, and in particular the linkage of NFC to SIM cards. This has had numerous side-effects:

  • It's delayed the whole thing through massive bureaucratic & political procrastination
  • It's created a technical structure which means that transactions are actually too slow on many phones (eg turnstiles on London's Tube need to work in 300ms from tap-to-open, so people can walk through without breaking stride. Oyster cards work, phones don't)
  • It's meant that NFC hasn't been properly opened up to developers as a general API to do cool stuff with.
  • It wouldn't be able to work well on non-SIM devices (eg tablets) and would likely have a hard time dealing with dual-SIM devices, or the half the planet which either has 2 phones, or swaps SIMs all the time
  • It's led to ridiculous protracted trials & consortium-forming which has earned a lot for lawyers and PR people and totally messed it up for everyone else.
But the simple fact is that the whole mobile payments thing is a chimera. Yes there are some corner cases (unbanked people in Africa on M-Pesa, specific apps like Starbucks, Square for accepting payments). But the basic notion of "paying for stuff with a mobile phone" is simply flawed. Firstly, cash & cards work perfectly well. I've never had a problem buying a sandwich & thought "what terrible experience taking £3 out of my pocket". I can use cards anywhere on the planet with a pretty good acceptance rate. Chip & pin means it's more secure than before. And I never see anyone bothering to tap their cards on the contactless readers either.

The idea of your purchases "going on your phone bill" completely ignores the fact that most people on the planet use PAYG prepaid and don't get a bill. Average outstanding prepay balance is something like $5, I believe. Most contract users won't want a sandwich or a flight on their phone bill - especially corporate expense managers. It just doesn't fit with our mental model of "phone bill", which many people don't both looking at anyway as they're on a standard plan. Linking purchases to credit cards stored virtually in your phone just seems pointlessly geeky & needs interruptive apps to be useful. I don't buy all this couponing & analytics hype either - it's just putting lipstick on the pig.

The idea of electronically transferring money without so much as a PIN or a signature scares me and most other people. I don't trust any of the parties involved except the card provider and my bank, and adding in the handset-maker and mobile operator just increases the already-too-high perceived risk. Note: this is totally different to my Tube Oyster card as that is stored-value & decoupled from my bak & credit accounts. The most I can lose is the £20 I top it up with. I can also use the card while I'm on the phone - I like to multitask when I'm travelling.

It's notable that the much-vaunted Japanese Felica system is still little-used for actual purchases of goods with phones. And that's despite NTT DoCoMo spending something like a billion dollars buying a bank and a stake in a convenience-store retailer to catalyse the market.

Schiller is right. The "tap-to-pay" thing is a nonsense, a solution looking for a problem. The involvement of a telco adds zero value and lots of friction. At some point I might want to use the phone to make transactions against a loyalty account (hence Starbucks), but that's likely to be very specific to a particular brand or store & I'd like to do it "in the app". QR codes (as used for airline mobile boarding passes) are not a bad option for this, and *maybe* NFC in the much longer term, but even then I still prefer the "visible" code - and no need for physical contact with the device. [I don't believe in phones for *reading* QR codes, but displaying seems OK]

Where the real value of NFC might lie (and I'm still not 100% convinced on these either) is in what I refer to as "interactions", not "transactions". Stuff like a "click to connect to WiFi" pad in a cafe, or a "touch-to-like" Facebook icon in a restaurant. The WiFi example has already been done by Blue Butterfly is much more elegant & sensible than the pointless and wrong-headed "seamless" approach suggested by some carrier-WiFi advocates. I wrote about this 18 months ago - the volume of free non-monetised NFC interactions will outstrip paid ones by orders of magnitude, like free apps in the AppStore. Operators probably won't want to be involved in that loop - and will likely slow down the developer app-creation process anyway.

We need to get rid of cellular operators from the NFC value chain, except as just another class of app developer. There won't be many transactions for them to take a cut from anyway, and their involvement in "interactions" just adds extra complexity and bureaucracy without providing any value. We also need to bin the idea of NFC being transactions-first entirely, as it has perverted the entire development course of the technology. It *might* come later, once normal uses of contactless have crossed the barrier of public acceptance - along with trust that it's OK to tap your precious device on things.

Apple has not "killed" NFC with iPhone 5. It's just merely pointed out that it's on its deathbed already, dying from a virulent payments & telco-involvement disease. It might be resuscitated, but I doubt it - and Apple has cleverly avoided contamination from its corpse.

One last thing: If a miracle happens & NFC does start to recover from its debilitating case of payments, then tablets will make great NFC *readers* for many applications. Apple, Google, MS, Samsung & co. should be embedding that functionality, before Square takes even more of the merchant market away from them.

8 comments:

  1. Spot-on. I've always thought the primary use case for NFC would be exchanging contact info with someone you've just met. We could do it seamlessly and in a cross-platform way using IrOBEX a decade ago, but somehow Bluetooth lost this basic capability, in practice if not theory, probably due to concerns about spamming. Bump.com, as implemented in the LinkedIn app, for instance, sort of works, but is really a hack.

    The big problem with NFC from an interaction design point of view is that touching something with your phone does not carry sufficent sematics to determine user intent. Thus the user has to interact with their phone and at that point they might as well use an app. There is still a need for proximity detection, and doing this using GPS as bump.com does is not ideal. Apple has an ad-hoc peer-to-peer solution for multiplayer games (up to 16 participants) that works seamlessly using either Bluetooth or LAN + Bonjour, and WiFi Direct would also allow that, if it were more widely implemented. No actual contact needed.

    ReplyDelete
    Replies
    1. strongly disagree. In reality, no one ever bothers with electronic contact info. Those are just trash information, to be lost in electronic address books. And worse, it's not even up to date, when you have lots of business contacts.


      A card will offer a tangible reminder, has space for notes and can be acted upon when found again. Even when im the decision is to just toss it. That's better than adding superfluous information to your computer.


      Delete
  2. I agree also iwith the premise of this piece, but mainly because much of the oise about NFC is payments related, therefore telco related. Many years back when Nokia came out with their NFC trial mobiles, they had a similar approach. The second time, they concentrated on just transport payments (bus terminals and such). Now, their NFC play is about triggers. You are playing music, and want it to go to the speakers, so you tap the mobile to the speaker you want to play it from. You have a Bluetooth headset and it isn't not yet paired to you mobile; instead of the Bluetooth paring exercise, you tap the devices and the pairing is done.

    If NFC is going to do anything, it will be these latter interactions, and the WiFi hotspot (trigger a utility) approach, not payments.

    ReplyDelete
  3. Anonymous7:01 pm

    The fact that you didn't consider the other benefit of NFC (wireless networking) is strange.

    ReplyDelete
  4. Anonymous2:58 pm

    If I got an itemized receipt in a usable electronic form I would probably use NFC for payments.

    ReplyDelete
  5. This comment has been removed by a blog administrator.

    ReplyDelete
  6. Some very valid points, but I don't agree on some points.
    For example, you mention "your purchases "going on your phone bill"": this is not related with NFC. While technically possible, I think it's unlikely: the payment feature would not be connected to your telephone bill.
    You also mention "cash & cards work perfectly well": very debatable. While Europe has deployed cards with a chip, making the transactions secure, this has not happened in the US, where users are still swiping the cards. This makes the card easily cloned. Someone is paying for that (mostly merchants due to high fees). Similarly, cash is less and less frequently in the pocket of people, due to large diffusion of cards.
    Last, "QR codes are not a bad option for this": no, they are not, but have two issues: they stop working when battery is empty (I don't want to walk because I can't pay the bus fare) and they are easily duplicated... in the sense that I can give the same QR code to anybody sending the screenshot, so this opens a bunch of new scenarios (there is no guarantee of authenticity and uniquely).

    But having said that, you are right on many of other topics.

    ReplyDelete
  7. "Apple's doubtless consumer-centric and design-first approach."

    What, like ios6 maps?

    I also disagree with Schiller's comment - I don't think we would progress particularly quickly if we only developed things that were a solution to a problem. It's a slightly restrictive standpoint.

    You've made your mind up and are so entrenched in your anti-nfc polemic that it doesn't come across you are thinking about it in a balanced way.

    Do you honestly believe that getting out your passbook, finding a code and holding it under a scanner is better than simply holding your phone next to a ready to redeem a ticket?

    You come across luddite, I'm sure you're not.

    ReplyDelete