Fantasy CTO - how would you duplicate Barclays iPad platform?

We know that Barclays have bought 8,500 iPad to build a new retail platform, but if you were given the task of doing the same what would you do?
Written by Matt Baxter-Reynolds, Contributor

There's a storm coming in. I know it's currently fashionable to frame the IT industry in terms of weather metaphors (*cough* cloud *cough*), but the storm I'm referring to is what happens when enterprises start using post-PC devices as a platform for delivering enterprise solutions.

We now know that Barclays are trying it. They've bought 8,500 iPads for their retail operation. And now Charles Arthur has pulled some of the threads together so we have a better idea of some of the technologies that are playing a role in this project.

But if you were Barclays retail COO Shaygan Kheradpir's counterpart at another bank, what would your final solution look like? And how much is it likely to cost?

To be clear in this article, I'm not saying this is what Barclays have done. Imagine this is a game of "fantasy football" -- if you were the CTO of a bank, what would you do?


Here's what we know. Barclays operate around 1,700 branches in the UK. That's an average of five devices per branch. (Typically you'd want to hold a small number of devices as spares, but the numbers that we roughly no seem to suggest that they're not.) We'll assume our fantasy bank has the same number of branches the same number of devices per.

(As an aside, that seems too small a number to me, almost implying that this project is still at quasi-pilot phase. Some Barclays branches are tiny, and some are beyond huge. Two would be the minimum number of units per branch given that you'd probably need one on charge all the time, and seeing as most branches will be smaller it suggests the whole estate needs to be larger.)

This project is not going to be a case of "wait for UPS to deliver devices to branch, unpack, and off you go". It's a good deal more complicated. As well as the devices we'll need to deal with getting power into them, stopping them from getting damaged, stopping them from going missing, data security, application delivery, user training, ongoing maintenance, and retirement. Three of these -- data security, application delivery, and ongoing maintenance -- are classic mobile device management (MDM) issues. The others require us to box clever.

The case

Chances are, if you buy an iPad personally, you'll also buy a case. The situation here is no different, but the reasons are. You probably bought a case to stop your iPad from getting damaged. For bank use, you might do it both to prevent damage and as a way of controlling theft.

If you consider the normal computer you'd get in a bank branch, it's not very appealing to thieves. I'd wager that they lose few desktop computers to petty theft. But the iPad is a different proposition -- iPads are highly desirable targets for larcenous youths. For US readers, the UK "High Street" (i.e. "Main Street") where these branches are located are often in deprived areas. Criminologically speaking what will tend to happen here is this sort of low-grade, petty theft is usually used to feed an addiction. The structure of UK towns tends to attract those engaged in criminal behaviours into the town centers, which happens to be where the bank branches are typically based. It's not a good mix if you want to keep hold of your assets.

The trick then is to make the iPad un-stealable, and the easiest way to do this is to make a custom case that you glue the iPad into. This makes it obvious at the local pawn shop that the rightful owner may not be the person standing in front of you by virtue of the bank's logo emblazoned on it. Remember that the iPad has no serviceable parts -- glue it in the case on inception and it'll stay there right until retirement.

That also means that bank staff can't take the case off before they inevitably drop them. And remember that the bank floor is usually not carpeted with a rich, deep pile. Plus a lot of staff will be using them when standing. Drop one without a good case at that height and it's goodnight Vienna. (Look what my ZDNet colleague Jason Perlow's wife managed to to do his.)

Let's be clear about the theft angle though. I'm profoundly relaxed about data security on these devices. the UK government now regards iOS 6 as suitable for data classifications of "restricted" and below. More to the point, it would be so insanely boneheaded for a bank to get this wrong -- well, it's not really worth discussing. The theft of an iPad from your fantasy bank should be nothing more than an irritation. It certainly shouldn't demand that your PR team switch into high gear.

Of course you can also do what you do in a normal retail store and have an alarm go off when one of the devices was taken outside of the store, but in a bank how would that work? A normal store has security guards -- in the UK banks do not. A member of bank staff is hardly going to feel inspired enough to run off down the street after an escaping iPad. I think we'll be able to tell how Barclays feels about theft by watching how keenly their staff hold onto them. If they treat them like a doting parent with a newborn babe that's one thing. If they treat them like yesterday's discarded newspaper, that's another. Training will instil one viewpoint or the other. But I digress.

What about the cost? For the cases we'll add £5 ($8) per device, or £40k ($64k). It's the smallest individual item we'll call out, but I've called it our primarily because someone needs to manage getting them made.

So, which unit we should buy. If you're someone like Barclay's there's no way you would buy the cheapest unit you could, which takes the 16GB units off of the table. Whatever we spend has has to be written off over a long period of time. Apple don't say when support for their devices will end, but Barclays has to be looking at support for multiple years. Let's say three. Five sounds better, but I'd suggest that was too long given how fast things change. 32GB gives more room to grow, especially if you want to put videos on the device which you might do in a retail setting.

That means we're looking at £479 ($760) per unit, or £4m ($6.5m) for the lot. And forget about a deep discount -- PR on this deal is good, but Apple traditionally don't discount.

But about about 3G? Well, we know that Wi-Fi will be available in the branches, and the likelihood is that this would be covered off in another budget. But what 3G gets you is GPS, and also a way to kill a device when it's outside of the coverage of your corporate Wi-Fi. And although most thieves when they "nick" an iPad or iPhone just rip the SIM out to prevent it from phoning home, remember we can glue our case over the SIM slot preventing such things. (It's amazing what IT problems you can fix with superglue.)

So do we want to take an additional £100 per unit charge for a cellular modem? (That's an additional £850k [$1.4m] charge across the the whole estate. By the way, that's a 20% premium per device taking our total to £4.9m [$7.9m].) And we'd have to pay to plug it into the cellular network. Banks have a good deal of leverage when it comes to negotiating telecoms deals. Perhaps £5 ($8) per device, per month to enlist it on the cellular network seems fair. That's £510k ($816k) per year.

This then is a tricky one. We likely can't just ignore the theft issue. If we pay the extra for GPS and cellular on the device we can proactively manage the estate and kill devices that we can't find. I think I'd like that peace of mind.


To actually use the device, you'll need to get some power into it. You won't see someone furtling under a desk for a proprietary power connection. What you'll have is a unit in the back office with some slots into which you slide your iPad. Perhaps you'll buy specialised charging station like this one. The company that makes them, Bretford, apparently charge around £600 ($1000) for such a thing. Other vendors will be available and you might be a able to bring the price down a little, but a) remember you're a bank and you're not scared of capex, and b) scrimping on this might result in killing the iPads through a buggy charging circuit which is embarrassing. That's £1.2m ($2.3m) at retail. Likelihood is you can get a fine discount on such a large purchase, so we'll take 20% off to give us £960k ($1.5m).

Remember though that you'll need to deal with issues like Apple changing the connector from time-to-time. And also remember to buy one that will accommodate whatever case you choose to put on it.


From Charles's article we know that Barclays has chosen MobileIron as it's MDM platform. Specifically Charles references that it used to provide an internal app store and also that's used for keeping track of where the device is and locking down what networks it's allowed to attach to. If you've never seen MDM it's essentially what you'd expect it to be -- it allows you to provision, upgrade, and kill off devices, and they all have this sort of "private app store" arrangement that allows whitelisting from the public app stores, and manage distribution of private line-of-business/sideloaded apps. You'll need to use that and the iPhone Configuration Tool (it's badly named, it does iPads) to set-up your baseline implementation.

The MDM story on iOS is now looking really solid. MobileIron is just one vendor -- AirWatch and Zenprise are others, and all three happen to be listed as "leaders" in a May 2012 Gartner "magic quadrant" report on MDM.

It doesn't matter which we choose, but we'll need to budget £2 ($3) per month per device, including any discount we might be get. That's £200k ($320k) per year. MDM getting better is just one reason why we may well now be on the cusp of a tipping point as far as enterprise adoption of iOS goes. You'll also need some implementation hours and some backend servers to run that on. Let's add a £25k ($40k) one-off charge for that.

That's also the first of our ongoing costs, and the easiest one to guess at.

The ongoing cost that's hard to guess at is the management cost. In fact, good luck getting any figures on this at all. Mobility total cost of ownership (TCO) figures are very hard to find, not least because it's relatively rare and very new. But someone has to pay for the help desk calls, the project management, the logistics -- all the stuff we have to deal when it comes to time and effort involved in managing kit within the enterprise.

PC TCO is better understood, and is always scarily high in comparison to the relatively cheap cost of the kit. So what do we do -- say that our TCO over three years is double the cost of the base units? That's slightly less than accepted wisdom of desktop management, and tablets are arguably easier to manage. I think that's likely a fair guess, so why don't we add £1.6m ($2.6m) per year by way of a management charge.


Finally then, you have to teach people how to use the things. This isn't to be taken lightly -- if you're buying 8,500 tablets you'll have 8,500 staff members to train. If you spend a day each training them, that's 23 person years on aggregate.

I don't have a good answer for the cost of training. There are way too many variables here. Let's just add £1m ($1.6m) as a training budget. Remember you're not just training someone to use an app, you're training an entirely new approach to how you do business face-to-face in the branch. That only works out to £117 ($188) per user.

Maintenance and retirement

I just need to cover off the topics of maintenance and retirement.

Over the life of the project, these things are going to break, get lost, get stolen, or just stop working. Our MDM platform should let us provision new devices. The only problem we'll have to deal with is oddities like hardware changes. We may have a new power connector, or buttons might move which will make our custom cases or charging stations obsolete.

Retirement is also easy -- once a device gets to the end of its life we'll hand it over to whichever corporate recycling partner we work with.

In summary

So what have we built? We've got a stack of 8,500 iPads. We have a software solution for provisioning them, keeping them secure, and pushing out apps. We have some protection against them being so readily stolen using relatively primitive protection of gluing them into case that's unattractive to thieves. Buy ponying up for 3G and GPS we can reach out and kill of any devices that we might have lost. We have a way of charging them neatly in the back office. Finally we have a way of teaching people how to use them.

How much then have we spent? Our up-front costs look like this:

  • 8,500 iPad v4, 32GB, cellular: £4.9m ($7.9m)
  • Cases: £40k ($64k)
  • Charging stations: £960k ($1.5m)
  • MDM servers and implementation: £25k ($40k)
  • Initial training: £1m ($1.6m)
  • Total up-front: £6.9m ($11m)

Our ongoing costs look like this:

  • MDM: £2 per device month, or £200k ($320k) pa
  • Cellular connectivity: £5 per device month, or £510k ($816k) pa
  • Internal management cost (helpdesk, PM, etc): £1.6m ($2.6m) pa
  • Total ongoing: £2.3m ($3.7m) per year

The grand total of our three year project? £13.8m ($22m).

And what do we have? A totally new post-PC platform for running a new generation of apps on providing an entirely new way of engaging with customers.

Worth it? Oh yes.

What do you think? Post a comment, or talk to me on Twitter: @mbrit.

Editorial standards