The weather stinks, wish you were here! The forecast proved 100% accurate for my latest tech conference, TechTomorrow in Columbus, Ohio: Rain, with a 100% chance for agreement that mobile's important, big data is hard, and damn it, we really, absolutely must get IT aligned with business goals.

Every event faces intermittent obviousness, but good ones help the clouds part so smart ideas shine through. Here are six useful things I heard.

"I need an IT partner in everything I do." That's a chief marketing officer--Shelly Stotzer, CMO and VP for publishing company Highlights for Children.

IT's latest existential worry is that its tech spending will be eclipsed by the CMO's as marketing becomes more data-driven. Meanwhile, CMOs just want to make sense of the loads of data they're accumulating, get mobile apps built faster, and know when social networks light up with flames of glory (and disgust) about their products. The problem is that too many IT teams have a strained or nonexistent relationship with marketing, so marketing doesn't even think of IT as a resource for tech-intensive problems.

Stotzer said a CMO needs someone who knows what's possible with technology and can help wring out a fix--not someone who waits for the CMO to request a project. "I need you to figure out what I really need," she said.

"A lot of times we ask, 'Why did the customer do that?'" Tara Paider, associate VP of IT architecture at Nationwide Insurance, said its statisticians look at its data to understand customer trends. But Nationwide also has a behavioral psychologist, along with marketing, tech, and product specialists, on a cross-discipline team.

While emerging tools like Hadoop let companies explore new unstructured data sources, Paider emphasized that companies already know a lot about their customers from conventional, structured data warehouses, and those environments aren't going away. Better decisions will come from combining the new insights from unstructured data with existing customer knowledge. "Looking at big data out of the context of the information you already have about your customers and policies and products is a bad idea," Paider warned.

"Digital is completely different from steel lift trucks." Mike Gallagher is VP of design at Crown Equipment, a maker of lift trucks that for the past seven years has been working to make technology a bigger part of its products. Its software includes tools that monitor data such as the operating efficiency of fork lifts--and the people who run them.

Crown has learned that designing a digital component takes far more collaboration than designing a mechanical part, Gallagher said. Software elements are different because they connect to so many parts of the product, and because the user interface is so central to the experience. Plus, they change much faster than mechanical parts.

Global CIOs: A Site Just For You Visit InformationWeek's Global CIO -- our online community and information resource for CIOs operating in the global economy.

As more machines are linked into an "Internet of things," more companies have to bring digital experiences into their products. They'll learn it's hard to mesh the faster, more collaborative software development cycle with an existing product development process.

"We refer to mobile not as a channel but as a metachannel." Risking obviousness here, but it's a point well said by Scott Gamble, VP at Alliance Data. A mobile strategy used to mean mobile marketing and commerce. Now it's understanding how people use a smartphone while shopping in a store, visiting a resort, filing a claim, or talking with a patient.

"Once you've spent a certain amount of money, pilots don't fail." Cloud infrastructure and software should mean we now do fast and cheap experiments rather than slow, costly pilot projects, said MIT's Michael Schrage (author of Who Do You Want Your Customers To Become?). Doing so would make it more likely for people to do real experiments--ones where the outcome's uncertain and it's really OK for them to fail. With cloud infrastructure, if an experiment works, the company can scale up more quickly. "The future of innovation is, how do we learn from small experiments and scale them?" Schrage said.

Run these experiments on "customers the organization respects." Also from Schrage. It strikes me that this is the part of cloud-based experiments that's not cheap and easy. The attention of your best customers is precious, and you can throw only so many half-baked ideas at them, yet it's the opinion of great customers that companies value most. So I see this as the main constraint on ramping up more cloud-based experiments. Schrage, however, argued that "your customers will think you look smart for approaching them."

InformationWeek: Oct. 15, 2012 Issue

Download a free PDF of InformationWeek magazine
(registration required)