Here’s the awesome and frustrating thing about covering startups. The second you think you have a rule of thumb for if they’ll work or not, you find an exception.

Remember the obsessive late 1990s need to be first to market? Well, then we had Facebook and Google and the iPhone and the Kindle– all better versions of social networks, search, smart phones and ereaders done well after competitors.

Okay, then, maybe you should never be first to market because the pioneer always gets the arrows in his back. Frequently true, but then again, Twitter and LinkedIn essentially defined new categories.

The rule of “you need a grownup CEO” was considered iron clad pre-2000. But it has since given way to the reverse, that a founder should always stay CEO. In truth, there are examples of both working.

Be lean! Be fat! Raise no money! Build a war chest while you can, because the markets could collapse! Back a young, first time CEO. No, back a serial entrepreneur. The reality is there are examples of success on all sides.

Even the golden rule of “Don’t ever try to change the music industry” has two examples that have (so far) bucked it in Pandora and Spotify.

But there’s one rule that I always remember, when I’m assessing the viability of a startup, one that I’ve seen very few companies defy: You have to design Web apps for the way people are, not the way they wish they could be.

That simple reality is why most health and weight loss apps and most money management apps have failed. People sign up, and they mean well. They genuinely want to change. But they don’t keep up with it for the same reason that they don’t change their habits in a non-Web world. And then, that site is just a reminder of how badly they’ve failed.

It’s like the advice your mom used to give you about dating: You can’t change him. (Or her.) Likewise, the best designed site in the world simply can’t make us someone we aren’t. The flaws of humanity simply aren’t a UI challenge waiting to be solved.

Then again, if you listen to the Obvious crew, perhaps changing humanity is just a really hard design challenge waiting to be solved.

Obvious’ latest company launched today. It’s called Lift and it hopes to make you just slightly — just ever so slightly! – better, allowing you to track goals and get props from your friends when you make progress on them. There’s not much to it– by design. But, according to Lift co-founder Tony Stubblebine, Evan Williams called it one the hardest, most subtle design challenges he’s ever seen a team grapple with.

Lift is a free mobile app that is drop-dead simple to use. (Tip: Search for “Lift Worldwide” or your iPhone will be flooded with apps to help you tone your butt.) You sign up. You add some things you’d like to work on. They range from drink more water to floss more to tell your wife you love her to stop buying things on Fab everyday. “Lift will work for anything if you can break it down into a routine or a behavior,” Stubblebine says.

The goals are mostly small steps towards a better life, which is great because small steps feel doable. And you don’t log, say, how much water you drank. You just hit a green check mark when you’ve made progress. What constitutes progress is up to you. It may be drinking a glass; it may be drinking the full allotment of eight glasses a day. The point is you did something. You can annotate it or ask questions and your friends can give you props. And that’s about it.

I decided to work on flossing, drinking more water, eating fewer sweets and exercising. After an hour of Zumba (you can laugh) during which I drank a gallon of water and then passed up some sweet vanilla yogurt, I got to check three things. And by the time I’m writing this Bijan Sabet had virtually high-fived me on the sweets. The hope is that little nudge keeps me from indulging in sugar bit-by-bit. My fear is the good feeling wears off.

There’s a forced social element to all of this. You are grouped according to goals with strangers and can interact separately with friends about these goals. You can’t be private: You can only be public or have a pseudonym. There was a lot of debate on that one, but Stubblebine felt that people always overly value privacy and underestimate the serendipitous value of being public on the social Web. But given that some people may not want their names attached to “drink less” or “spend more time with my kids,” he admits that will probably change over time.

There are some remarkable ripple effects from forcing it to be open. For one thing it’s very humanizing to see well known entrepreneurs, whom we tend to put on pedestals, struggling with things as common as remembering to floss.

The benefit of the friend circle isn’t just motivation but tips from other people going through similar struggles. For instance, when Stubblebine tried to go vegan, he used the app to ask vegan coworkers where they eat lunch. Other examples might include tips like drinking water instead of soda for meals, or using a special luxurious cotton floss that could increase the likelihood of flossing.

Unlike a lot of fitness apps, Lift decided to be a total nag-free zone. There are no reminders or alerts to suck you back into the system. That was a controversial move, rooted wisely in how people actually are, not how they’d like to be. “People were always telling us they needed reminders and when I asked if they use them now, they would say, ‘Yeah I have them and I ignore them,'” Stubblebine says. “It’s clear nagging doesn’t work.” Instead, Stubblebine hopes the positive reinforcement of actually doing things and wanting more of that will become addictive.

Lift is the latest company to come out of Obvious, the incubator or accelerator or fund or company generating machine or whatever they are calling it these days started by Twitter co-founders Williams and Biz Stone. Other investments have included Branch, Medium, and Karma.

As Erin discussed last week, Obvious seeks to build the next Wikipedia not the next Zynga. Put another — and more overused — way, it wants to make the world a better place. I know, you’ve heard that one before. But you can actually see it with most of Obvious’ companies.

Branch seeks to bring order and civility to one of the greatest unscalable cesspools of the Web: communities and conversations. Another Obvious investment, Karma, actually did change behavior in my case: It made me a better gift giver by making gifting creative, super-mobile, and drop dead easy. It enables that “oh shit!” moment on the subway when you remember your mom’s birthday and a few clicks later, you’ve sent a gift. No more sending yourself an email or writing yourself a note on your hand to remember.

There’s less written about Beyond Meat, which makes uncannily realistic vegetarian meat products now being sold in Whole Foods and elsewhere. As a Southerner, it’s hard for me to get on board with that one, but, hey, good for someone else.

I’m sure Medium fits in here somewhere, but I still can’t really figure out what it is, other than — I’m told by other bloggers — the future of publishing. One of Obvious’ more annoying ways of changing the way startups are born is its habit of just putting up their own blog posts on product and company launches, rather than working with journalists on stories. That gives us little more than launch posts that describe a new and mostly closed product like Medium in vague catch-all terms — like mostly “vision,” “putting users first,” only now “a sliver of what it could be” and an “evolutionary leap.” (Perhaps the hope is to appeal to people who love the simplicity of Tumblr but wish it could be slightly more insufferably, hipsterly, opaque.)

To be fair to Obvious, these values aren’t just luxuries adopted later in life by two guys who’ve already made it. Williams’ two hits, Blogger and Twitter, were filled with “make the world a more open democratized place” warm-fuzziness.

The timing is good for Lift, as mobile devices seem to be doing a better job than traditional websites at changing our bad habits. The key to an electronic pat on the back is being able to capture a milestone the moment you have one and entering that data as seamlessly as possible. Consider the impact of things like the Nike Fuel Band or the Up, which travel with you and integrate seamlessly into Path — in the case of Nike– where your accomplishments aren’t overshared the way those annoying self-Tweeting scales worked.

“The ideals are all clearly there. We just have to show that these companies can work,” Stubblebine says.