Every Friday afternoon at my startup, Iodine, we gather in the conference room over beers for show-and-tell, spending an hour or so catching up on what the team has accomplished over the past week.

For many months, show-and-tell was basically a brisk demo session, during which our designers and developers would run through slick new interactive features, clever engineering hacks, or tempting mockups of upcoming products.

These days, show-and-tell is less about prototypes and more about spreadsheets. Where we used to “ooh” and “aah” at nifty app tricks, we are now drilling into the numbers, facing the sometimes stark truth of what’s working and what’s not.

The objective of these sessions is to expose last week’s hunches to the bright light of today’s reality. Soon, we’ll be making a case for more funding to keep Iodine growing, and we’ll need a story to tell, a story nestled somewhere in all those metrics. We need to mine those metrics until the story emerges.

Of course, different numbers tell different stories, and part of the exercise is to discern what story certain numbers tell and what story we want to tell, and hope that the two converge. So it’s been clarifying to learn, for instance, that week-over-week user growth--the conventional “metric that matters” for consumer Internet startups--isn’t necessarily the story we want to tell, nor the one that will lead us toward our goal. Yes, we’re building digital consumer products, but our revenue model won’t depend on advertisers, which makes chasing eyeballs alone a blind alley.

Rather, the story we want to tell starts with how well we’re converting the masses into a devoted legion. We want people to find us from a Google search or Bing ads or Facebook shares, and leave feeling better informed, more assured, and compelled to share their experience. In particular, we want to convert visitors into contributors, people willing to post their stories and experiences with medicine (and medicines). Out of that metric, we can spin not just a story but also a revenue model, of individual experiences creating a new data set, one that informs a way to optimize medical therapies. In this lies the prospect of something truly valuable for our users and for medicine.

That’s a sweet sentiment, but sentiments can fade in front of hard math. There’s nothing more sobering than confronting the evidence that a Me Too button that we were certain would prove irresistible to our users has earned just a few hundred meaningful interactions. Or that a personalization engine that we figured was something close to magic is simply overlooked.

But minding the math can also turn failures into wins. When we looked at how many people were completing a medication survey form, we wondered if we could improve the numbers by adding nine words of text explaining why we’re asking people to pitch in. Presto--the conversion rate doubled. Realizing such dividends affirms the work that the math requires, and it demonstrates how much math is left to work out. The product we spent months proto­typing and building and celebrating has now become a robust platform for experimentation.

Really, that’s been the biggest lesson of our Friday afternoons. Show-and-tell has become an occasion not just for the flashy stuff but also to hail the rigor that we expect of ourselves. True, numbers aren’t as fun as shiny new product features. But it’s been especially gratifying to see how ably our small team has made the transition to the spreadsheet era. After all, the small stuff may be the stuff that, in the end, lets us tell our real, best story.