Gone are the days when traditional software companies could do two years of research and, only after all the intricate information is collated, would they set about building a product.

If they did that today, the product they would end up with would already be out-of-date by the time it hit the market.

Today, innovation starts with software development teams being able to understand data at a deep level to predict what may happen in the future.

Building new solutions happens in small iterative stages so each new element can be tested for market fit and fine tuned, usually while the platform is live.

>See also: 5 indispensable tools for Big Data visualisation

This means modern software-as-a-service (SaaS) companies are now constantly updating their offering, such as with Apple’s iOS or Xero’s cloud accounting platform.

Updates are continuously being pushed through in the background – and, with software being hosted in the cloud, users don’t get charged for the privilege of receiving a new feature or having a bug ironed out.

Customer usage: the key to progress

In the old days, businesses would sit behind a pane of glass and watch someone interview their users. Today, data around how customers use the product is the new focus group.

SaaS companies have infinite amounts of data that can inform what customers are actually looking for. Data informs what is built tomorrow and how the current product can be improved.

SaaS companies don’t necessarily spend millions of dollars on deep research – the research is often based on how customers use the product today.

Deep usage studies can be conducted with groups of customers, but with millions of small businesses in the world, that doesn't scale. A well-built SaaS service is designed from the ground up to measure customer behaviours.

Many customers will say what they want, and that’s important when it comes to the current product, but many can’t say what they need until it’s put in front of them.

Harnessing the power of machine learning

Machine learning was once only available to big business but now it’s everywhere – even in smartphone users’ pockets with personal assistant apps like Apple’s Siri.

Machine learning for small businesses should be around automating tasks and processes so small business owners have more time to do what they love.

As systems get smarter there will be more processes that can be automated in every area of business. Systems will even get smarter at anticipating what customers want.

>See also: How to measure the value of big data

And when it comes to notifications, they shouldn’t overwhelm users or drive them crazy. We live in a 1.0 world where you would have to spend an hour to lockdown your notifications, so you only see the ones you care about.

In the future, personal devices will be smart enough to understand the notifications users care about, so they won’t have to turn them on or off for each individual app. They will be based on users’ actions and how they interact with the system.

Small businesses don’t need flying robots, Star Trek or the Jetsons, but a system that is intuitive enough to predict their next action.

The only way SaaS companies are going to deliver that is by analysing user data to extract useful insights which drive product development – giving businesses the freedom to work smarter, not harder.

Sourced from Angus Norton, chief product officer, Xero