Web Applications or Desktop Applications, Which are Better?

What should be on your desktop? And what should be on the web? Does anyone really know yet? Is it even something you can generalize, or is the desktop/web debate really coming down to personal choice. Or are we in fact grappling with a new model of work?

Take email. Please. Email is still the lifeblood of most internet based communication. But now you have a choice among using email in the classic desktop sense (ie. tethered to your desktop), in the mobile sense (ie. on a mobile computer) or in the virtual sense (online email like gmail). Is the world better if everyone converts their email to gmail. It's certainly better for google, and depending on your usage patterns, it might be better for you, but its simply not true that hosted email is better for everyone.

When my mother comes to town, she likes to keep track of what's going on in her life. Even though she's retired, she's busy with a social life and a philanthropic life (helping to run a 10,000 villages store), and the ability for her to check up on things even while she's in Seattle visiting the grandkids is perfect for her. Gmail is a simple user interface available over just about any browser and it lets her read and respond to email. It's a great solution for her.

So why doesn't this work for corporations? Two reasons: 1) Requirements are heavier (like managing user accounts with actual employees) and 2) Email retention policies (I'm using retention in the sense of deleting email more than 1 year old). GMail is great for mother, not so great for big business. Could online email work for corporations? From a user interface perspective, yes. From a data management perspective, no.

But this pattern isn't the strict domain of our virtual lives. We see the same choices in our daily lives as well. For instance, what if I told you I could offer a guaranteed way to get you to any part of the city in short order for $1.50? You'd be interested until I lifted the curtain and showed you A BUS. At which point you'd say 'No Thanks, I have MY CAR.' Without twisting the analogy too far, the bus reflects a decision to use a 'common' email solution, like GMail, while the car reflects a decision to use a 'private' email solution, like Outlook.

What can the bus do for you? Well, in addition to getting you places, its really about what you don't have to do. That is, you don't have to manage the bus. Someone else-- the city or county-- performs the maintenance, cleans the interior and provides a driver. They even create routes, schedules and stations for you. And not only that, but someday's a new bus even shows up and you never had to talk to a bus salesman! All for $1.50 a trip, which, with gas over $3 a gallon, means the bus is the ultimate solution that everyone should use.

Except most people choose to buy their own car, which makes no sense economically. Could it be that people like the management aspect of owning your own car? The trips to the service shop for oil changes and repairs? Keeping the interior clean in spite of the two kids and a dog? Paying thousands and thousands a year to feed the beast and insure it? This doesn't bring a smile to my face.

But what does is being able to crank the stereo. Let's see you do that on a bus! Or maybe I want to change my climate zone to 66 degrees while my passenger's climate holds steady on 70. You can't do that on a bus (you can't really do that in a car either, but its a nice sales gimmick).

This leaves us with one big reason why we want our own car instead of sharing a bus: indulgence-- which we will rephrase as 'convenience.' This is why there are 3,478,923,241 features in Microsoft Office-- cause every one of us wants to be in charge of some knob somewhere. In fact, this is why car makers show thier wares in motion on empty winding roads with nobody around. The message is 'you're in charge of where you want to go.' (Kind of like 'Where do you want to go today' eh?) It wouldn't be as compelling if they showed the reality of me in my car putt-putting bumper to bumper with the masses on 405.

I'm not saying this is bad, just observing human behavior. Because we can still make rational decisions on our transportation, and do so every day when we fly in a plane. While its true some people can afford the management of airplanes, most people find fair value in the $500 ticket to fly somewhere and back and not worry about changing the oil in the jet engines. Just make sure they provide the plane comes with air flow knobs above your head for you to twist. It may not do a whole lot while you're stuck on the tarmac, but it feels good just knowing its there, eh?

So, back to software. How does any of the above reflect whether a user will choose a web application of a desktop application, or vice versa? There are at least 3 metrics at work:

1) Operation -- What is the perceived value of the app.
2) Management -- How important or necessary is management of the app.
3) 'Convenience' -- How much does the app make me feel like I'm in charge.

Let's give each metric a weighting of 1 to 10, with 1 meaning the least and 10 meaning the most. My mother might measure her decision for GMail vs. Outlook as such:

GMail
Operation (5)
Management (8)
'Convenience' (5)
-----------------
Total (18)


Outlook
Operation (5)
Management (3)
'Convenience' (5)
-----------------
Total (13)

Operation for my mother is simple emailing, nothing very esoteric. Management is heavily weighted towards GMail because she doesn't have to worry so much about spam or 'Auto Archiving'. Because my mother's use of email is more utilitarian, the extra bells and whistles of Outlook don't mean as much to her, so GMail wins.

How about the corporation with 1000 employees and the full time IT staff? They might weigh things as such:

GMail
Operation (4)
Management (4)
'Convenience' (3)
-----------------
Total (11)

Outlook
Operation (8)
Management (7)
'Convenience' (8)
-----------------
Total (25)

For the corporation, email is a very important application and the numbers above reflect this. Operation is better with Outlook because it may be hooked to an exchange server with active directory or an IMAP server for people on the road that require secure connections. And you get wizbang features like shared calendars for scheduling meetings. A corporation has a dedicated IT staff so with real IT people, they'll have a much easier job managing apps (and 'retaining' data) on their servers than on Google's. Finally, people will like some of the features like synchronizing their Outlook with their Blackberries and cell phones, adding personalized layouts and signatures and any of the other 1,000,000 knobs they might find.

How about a more complex application? Like Project Management for instance. Full disclaimer, I build project management applications for a living here, so I am very biased. But I still had to make a decision on how to build them, and here is a little window in to the process.

Web Based PM
Operation (6)
Management (7)
'Convenience' (3)
------------
Total (16)

TeamDirection Project
Operation (8)
Management (6)
'Convenience' (8)
-----------------
Total (22)

And I can justify every one of these numbers. But I have the distinct feeling it will take me another blog post to do so. The short summary is people recognized the ease of web application distribution for what is a group activity (monitoring and executing projects) but are just now starting to realize how desktop applications, which have always had the benefit of richer UIs for complex actions, can improve the management and convenience aspects for their projects. The trick is how to provide the knobs but still enforce (with style, dignity and grace) the rules for group activities. I'll go over our tricks, and other applications tricks, that are beginning to level the playing field once again.