When you think of manufacturing companies, you probably picture conveyor belts, milling machines, spot welders, and a whole slew of stamping, polishing, extracting and reformulating processes going on at the factory. But what you also should think about is all the darn data around the products these companies produce that resides in different files and across multiple applications and databases, and how navigating and presenting it in a way to get some meaning out of it can seem as complicated a process as operating a gear cutting machine or taking on a precision grinder.

That’s where startup Inforbix thinks product data semantics will make a  difference. The company was co-founded last year by CEO Oleg Shilovitsky and CIO Anatoly Savin to answer the problem of how to retrieve product data located in different places in manufacturing companies that affects development, supply chains and manufacturing systems, and it’s expected to bring a beta solution to light in the next couple of months.

What Shilovitsky sees as an opening for semantics in manufacturing is a more vertical specialization for particular domain issues around data integration. “Thinking of manufacturing companies, realistically they implement and use different systems. And companies that try to provide the ultimate data management solution for manufacturing companies fail today because to make this change, to transfer dozens or hundreds of data sources and systems in a manufacturing company, is not feasible,” he says. “The cost of change is so high, and the amount of investment that the company already has made in data systems also is very high.”

Consider how long, for example, it takes to build an airplane and all the heavy and expensive investments  – in CAD and Office files, emails, databases, enterprise applications (CRM, ERP, PDM, PLM) and homegrown systems –.for these projects that can stretch over years. “To go with one new system or yet another data management system to replace all they have is not only not feasible from the implementation side, but the cost of replacement can’t be justified by any business. There’s tens and hundreds of millions of dollars in data infrastructure,” he says.

Taking aim at the more granular data problems is its focus, using what’s already in place. One customer testing its technology, for instance, was experience a leakage problem with pumps they made. Data around the pumps was dispersed through the engineering (design information), manufacturing (lot numbers and factory processing routines) and supplier data silos. “When they know what is the manufacturing lot, what components came from particular suppliers for this particular design, when they combine all that together they can discover the source of the problem,” he says. “They can discover kind of a pattern to understand that the problem was with one of the four suppliers. If you understand the real semantics of product data you can do it – combine pieces and discover patterns.”

Of course, this could have been accomplished before – the fact that there are other avenues to integrating data isn’t the point. “The main point is how costly is your solution and the problem is how easy can you implement this solution,” according to Shilovitsky. “Someone can say it’s a typical BI problem. But to make a report takes two weeks for the average IT people in a company, and it’s expensive. … Semantic technologies can provide a differentiator in terms of cost. But it needs to be proven.”

And doing it in a SaaS model makes capabilities even more accessible to a significant part of the manufacturing base – the 70 to 80 percent of manufacturing companies in the U.S. that are small businesses, under 50 people. “Life is not just Boeing and GM –there are lots of small companies and we try to focus how to provide solutions for them, when they can’t hire expensive IT and they just use simple systems like Excel,” he says. That seems an inexpensive way to manage data at the start, but soon becomes clear that running after the data in those documents creates complications and costs.