Big data is growing rapidly, and not just among big corporations.
Data allows companies, both big and small, to improve their offerings and make more informed decisions — and data-driven companies are more valuable in both perception and reality.
Sharran Srivatsaa recently told me how his company, Teles Properties, uses data intelligence for its brokerage model. By mining, spotting and visualizing patterns in hyperlocal micro-market data, Teles accurately predicts likely sales prices for clients and properties. The firm has given its real-estate agents a distinct market advantage and sold homes for the highest negotiated prices in the shortest times among its competitive group.
And then there ride-sharing sensation Uber, which has used big data to disrupt an entire industry. The app relies on regression analysis to determine which neighborhoods will be the busiest and activate “surge pricing” to get more drivers on the road. Uber uses data as both a competitive advantage and a product.Earlier this year, Uber agreed to sell its data on customer travel patterns, joining the growing ranks of companies using data as a source of revenue.
Even an established firm such as GE is flexing some entrepreneurial muscle when it comes to big data. GE’s Predix software, which aims to integrate GE sensors to create a true Internet of Things, is capable of sensing maintenance needs, predicting breakdowns and sending performance data to R&D to rapid-cycle product enhancements.
The company is staking its claim in the age of data industrialization, and its price/earnings ratio continues to climb. That’s the perceived value entrepreneurs can also gain through big data.
Here are a few tips to start building a data-driven company:
1. Identify your data customers.
Data customers aren’t necessarily a startup’s customers. Uber’s data customer — other travel firms — is entirely different from the general public using its service. Video game distributor Zynga is actually much more, leveraging data from each game interaction and selling insights to determine which games users play so similar games can be created.
2. Find out what data they need.
Which insights will have a direct impact on customers’ daily actions, and how will that information be gathered? Will it be structured so it can be analyzed immediately, or does it need to be cleaned? Data is nothing without context, so entrepreneurs must translate it so it makes sense to their customers.
3. Build or buy data.
Once the data need has been determined, build the infrastructure to collect it or pay a third party to extract it. A data ecosystem can be built cheaply with Amazon Web Services, but a data scientist still must examine the extracted knowledge.
I’ve heard a lot of protests about the complexity and cost of starting a big-data project. The human power is costly; rather than set up and maintain unique servers, however, one can order the precise server specs needed in the cloud, where data is sent back and forth very easily. I don’t have a stake in Amazon, but if setting up a data infrastructure can be as easy as buying a book, it makes sense for everyone.
4. Emphasize the visual.
Data is science, but its visualization is an art. To make data actionable, present it in a way that’s both human and persuasive. Nate Silver, founder of poll aggregator FiveThirtyEight, is a pioneer of data visualization. FiveThirtyEight used statistical models to predict the outcome of the 2008 presidential election and has demonstrated the emotional appeal of data visualization.
5. Automate the product.
If the data collected is a product in itself, automate the collection of the input data and the delivery of the output. Think of an API as a software USB port, an interface to transfer data. Once code is set up to port data within a predictive model, one can also automate that model’s visualization and capitalize on it as a moneymaker.