Data analytics for the smart grid tends to come in two different flavors. In one corner, there are the big, expensive deployments from IT vendors like Oracle, IBM, EMC, Teradata, and of course, the startup that would be the energy data analytics king, C3 Energy. In the other corner, there are the remaining startups, which may offer only a fraction of what the big boys promise, but also come at a fraction of the price.

At least, that’s the general understanding of how the still-nascent utility data analytics market is developing. It’s hard to know for sure, because most utilities aren’t disclosing how much they’re paying for the latest large-scale deployments, although they’re quick to announce how much value they’re expecting to get in return.

That makes any information about comparative pricing worth hunting down. One such glimpse comes from California municipal utility Glendale Water & Power, which has publicly disclosed a February report (PDF) that shows a serious price difference between big contenders Oracle and C3 and the winning vendor, Escondido, Calif.-based startup Detectent.

Simply put, there’s an order of magnitude difference between Detectent’s $280,300 winning bid, and the multi-million-dollar proposals from Oracle and C3. Smart Utility Systems (SUS), the final bidder, came in at $427,716, still well below its mainline competitors.

Of course, Glendale has its own specific needs, and some specific financial constraints to meet, that may not apply to other utilities. Specifically,  Glendale’s 2010 smart meter grant from the Department of Energy allows for it to spend $181,000 on “meter data analytics software, implementation, and ongoing support,” which with a DOE match would only add up to $362,000 or so to spend on any solution.

Perhaps the utility would have chosen a more expensive system if it had access to the money. Oracle and C3 have a number of utility customers using their products, mostly big investor-owned utilities with big budgets and large-scale data challenges. But there are plenty of utilities like Glendale -- desperate for software to help them manage their new smart ecosystems, starting with smart meters and working their way on up, without the upfront cost of an IT infrastructure.

Meanwhile, Glendale has a lot of uses in mind for its new analytics platform, starting with revenue assurance, otherwise known as theft detection. The utility’s losses have averaged 10 percent a year over the past decade, compared to a U.S. average of 4 percent, and it’s looking for a system that can help it find and stop that non-revenue-generating energy from leaking away. Theft detection is an integral part of many analytics offerings on hand from individual smart meter vendors, who are working with partners, as well as acquiring startups, to bolster their offerings.

Glendale’s list of particulars for its analytics project also included meter performance and maintenance diagnostics, which sounds like the kind of network management challenge being tackled by the likes of Bit Stew, Cisco and GridMaven. It also includes customer service and call center operations integration, and customer outreach that includes peer-to-peer energy use comparisons derived from the data. That kind of customer-facing analytics is where companies like AutoGrid, Tendril and Opower are making their mark.

Eventually, Glendale wants to support “load disaggregation,” or data analysis that can tease out energy use of individual equipment and appliances in buildings from smart meter data. Startups like Bidgely, PlotWatt and Navetas are competing with tech vendors including Belkin and Intel to deliver disaggregation technologies today, and the spread of networked devices in homes and places of business will only add to the richness of that previously untapped data treasure for utilities.

All of these features might best be delivered as part of a big, overarching integrated platform of the type C3 CEO Tom Siebel promises. But financial pressures are likely to push more and more utilities toward a la carte, piecemeal deployments -- at least until the first wave of projects start to prove whether they're actually delivering the valuable insights they've promised.