Blog Post

California’s Smart Meter Battle: Google vs. Utilities

Stay on Top of Enterprise Technology Trends

Get updates impacting your industry from our GigaOm Research Community
Join the Community!

Updated with comments from Google: There’s a battle looming in California over smart meters and energy prices. Google (s GOOG) says the state should require its big utilities to give near real-time pricing information to every smart meter-enabled customer by the end of next year. California’s big three utilities — Pacific Gas & Electric, Southern California Edison, and San Diego Gas & Electric — have raised plenty of objections to that deadline, and the California Public Utilities Commission is holding a workshop in San Francisco on Friday to talk about it.

The debate, which could influence smart grid policies across the country, underscores an important difference between the two things Google wants utilities to provide — energy “usage” data versus “pricing information.” Electricity usage is a real thing that can be measured in real time with magnets and wires, either by a smart meter or lots of other devices. Electricity prices, on the other hand, are contrived, during or after the fact, by a convoluted market that has to keep demand and supply perfectly balanced at all times. Delivering pricing data in real time will be challenging for smart meter networks as they’re currently being deployed. So in other words, for utilities, delivering power comes first, figuring out who pays for it (and how much) comes later.

Update: Google has responded to this story, noting that it was not the originator of the CPUC proposal to ask utilities to provide near real-time data to smart meter customers. Rather, Google is one of many parties supporting that proposal which originated from the CPUC last year. At the same time, Google has been a consistent proponent of giving customers access to their energy usage data, says Michael Terrell, Google’s policy counsel.

Most utility customers pay steady, regulated rates, and don’t get to see these complex price fluctuations — at least, not yet. But even getting slightly more complex tiered or time-of-use prices to customers through their smart meters could be problematic for current utility networks, given that most smart meter deployments today aren’t set up to handle that. As Lee Krevat, director of smart grid initiatives at San Diego Gas & Electric, put it in an interview this week, “We didn’t put in an Internet to each meter, or broadband to each meter — and ‘real time’ really implies broadband to give near real time pricing data.”

Most smart meter networks, including those being deployed by California’s big utilities, are lower-bandwidth and designed to be read every 15 minutes or hourly, not in real time. While there are ways to get faster or more current price information to homeowners, Krevat doesn’t see such a network being the best, or most cost-effective, way, to do it.

After all, “The rates exist on our Web site. The rate schedule doesn’t change very often,” he said. “Do you want to spend your bandwidth transmitting something that could be figured out at a customer end point based on their consumption data?”

Or, to put it another way, would utility customers support paying for the ability to see pricing data? The customers are the ones who pay for utilities’ smart meter system upgrades through increased rates. That certainly differentiates the utilities’ incentives from Google, which wants usage and pricing data opened to third party systems like its PowerMeter home energy management platform. Google promises PowerMeter will be free, but building a system that can provide it with data may still cost customers in one way or another.

SDG&E is working with Google’s PowerMeter and has about 125 customers testing it out — but right now they’re using day-old energy usage information, and currently PowerMeter doesn’t deliver any real-time pricing information. Eventually, California’s three big utilities plan to turn on their smart meters’ home area network (HAN) connections, but they’re doing a lot of testing first. Krevat said that’s an important first step in designing a system that’s both cheap and effective — “Understanding the model for how the customer wants to use it is the first step,” he said. “Then you can decide the technical solution.”

Ted Reguly, SDG&E’s smart meter program director, said customers mainly want some kind of current bill calculation, as well as some kind of pre-set alert when that monthly tally gets too high. Someday people will want to hook up smart appliances and other in-home energy controls to the smart meter via the HAN. But as SDG&E noted in its comments to the CPUC filed in March, “the Smart Meter system as currently designed requires more than HAN to provide customer access to near real time information on prices.”

Beyond these issues, it will be important to clarify what Google means by “pricing information,” Reguly said. Does Google mean the flat rates homeowners are scheduled to pay, or the actual prices that they end up paying after the bill is finalized? “You might think the cost of electricity is X, but it’s really Y because of bill settlement two or three days later,” he said — and getting the more accurate figures to customers in real time would require utilities to completely overhaul the batch processing-based back-office billing systems they now use.

Andy Tang, PG&E’s smart grid chief, said during a recent energy symposium in Berkeley, Calif. that asking utilities to replace their batch-based systems with real-time systems was “impossible” in such a short timeframe, at least not at costs that regulators would be willing to pass on to customers. Tang also expressed some frustration with Google’s push for deadlines for delivering real-time pricing, given that the federal government is still working on standards for all the smart grid systems to make this possible, he said. As PG&E wrote in its comments to the CPUC, “No amount of cajoling or wishing by one vendor or another that it happen by an arbitrary date can change the need for development of such uniform standards.”

Emerging Standards

Just how those standards will emerge remains to be seen. ZigBee, the wireless technology that’s taken a lead in smart meter-HAN connectivity, is working on a second iteration of its Smart Energy Profile specification for energy data that will include some pricing information, Reguly said. For commercial and industrial customers, open demand response technologies like Lawrence Berkeley National Laboratory’s OpenADR or EnerNOC’s (s ENOC) PowerTalk, are expected to embed price signals as part of an automated system to turn down devices to help utilities reduce peak loads.

Perhaps broadband could be the solution. The National Institute of Standards and Technology, the federal entity setting smart grid standards, has asked the smart grid industry to comment on whether some or all of the customer’s smart grid connections should come through broadband connections independent of the smart meter. There’s a long list of companies looking at selling energy monitoring gear directly to consumers, either as stand-alone products or bundled with home broadband offerings or security systems. Google is working with utilities and smart meter maker Itron (S ITRI), but is also partnering with in-home energy devices from Energy Inc. and AlertMe with its PowerMeter.

All three California utilities have asked CPUC to avoid any hard deadlines in favor of looser policy guidance. But the issue COULD comE to every state. The Federal Communications Commission’s new U.S. National Broadband Plan includes some strong words for state utility regulators to encourage utilities to deliver real time pricing data to consumers.
To wit:

“States should require electric utilities to provide consumers access to, and control of, their own digital energy information, including real-time information from smart meters and historical consumption, price and bill data over the Internet. If states fail to develop reasonable policies over the next 18 months, Congress should consider national legislation to cover consumer privacy and the accessibility of energy data.”

Just how the CPUC decides to take up Google’s its deadline — as well as how it comes to define pricing data in the process — will be closely watched topics in the smart grid industry. Stay tuned for more details later this week.

24 Responses to “California’s Smart Meter Battle: Google vs. Utilities”

  1. Niall McShane

    Interesting article and discussion. The implication that real time pricing is dependent on billing settlement that takes place a couple of days after the actual billing interval closes implies a level of precision that is largely unnecessary. For years, the utilities have absorbed the fluctuations in generation costs within a single fixed rate billing structure. There does not have to be a direct algorithmic link between real time generation costs and real time billing.

    If we consider the telecom business, over the last several years they have moved away from a real time billing model to a model of monthly packages and, in many cases unlimited usage. Clearly, unlimited usage is not a suitable model for energy usage since the objective of the smart grid initiative is to reduce consumption not to promote more consumption. However, as with the telecom example, I believe that the utility industry can and should develop various pricing models that abstract usage to a higher level and allow for real time pricing. Some examples of possible models would include:

    • Monthly usage packages similar to those employed by the telecom industry. In these models, there could be several tiers of usage with each tier having a different per kWH price. Reduced consumption could be encouraged by charging premium prices for consumption above a specified level. An individual plan could have one or more tiers with the first N kWH being charged at the lowest base rate, additional tiers charged at progressively higher rates and any usage above those tiers charged at a significant premium. To be effective, the lowest tier would have to offer a significant discount over the current fixed rates. In this case, the pricing signals that would be sent to the consumer would indicate their current consumption against their monthly plan and could even predict the total bill based on consumption trends. Warnings could be sent when a tier threshold is crossed. The home energy management system could have a role to play here by maintaining data on average daily consumption and using this to forecast total usage for the current billing cycle. This would not be exact but would provide a signal that the consumer could act on and, over time, the consumer would get a feeling for how accurate the projections are. In fact the intelligence could be built into the systems to allow them to self calibrate over a period of time to become increasingly accurate.

    • Variable rates based on historical data. The utilities could develop models of demand based on historical data that predict, for a given set of attributes such as; time of day, day of the week, temperature, humidity, air quality etc, what mix of generation sources they need to call upon to meet demand and the actual cost of that particular mix of generating assets. Based on these models, they could derive estimated price points. These could be discrete based on certain thresholds or continuously variable as appropriate. These price signals would then be sent to consumers or to their smart meters/ home energy management systems to effect demand response.

    Other models are possible too. The important point is that moving to delivery of real time pricing does not necessarily mean that the utilities need to completely overhaul their current offline billing systems and make those real time.

  2. Jon Husen

    Just to elaborate on a few of the points made. The first question that pops into my mind when thinking about real-time pricing is “what is the value-added benefit of this feature?” Under current pricing models, The basic billing formula is simple arithmetic -> kWh x rate + surcharges = your bill. This feature, at least in its inception, will be either an added benefit charged to the customer for a nominal fee or a feature for everyone with the cost thrown in with the other surcharges. Do customers really want to pay for that feature when it’s so easy to figure out on their own?
    As the smart grid infrastructure continues to develop and dynamic pricing models come into play, there will be a greater benefit to having real-time pricing; however, that benefit is still dependent of which model is employed. A simple peak/off-peak rate plan will give two different costs of electricity. The effect on the basic billing formula is minor. Truly dynamic, real-time consumer pricing would require significant changes to how utilities handle power purchases, billing, and network design. The purchase of peak power on the open market injects a mess into the pricing algorithm. While this can be sorted out, one must ponder “at what cost?” When considering what information would result in customers making meaningful changes to their routines to save electricity, do they need to have an in-home display with a continually fluctuating cost of electricity or would a simple message with the current and/or previous day’s total cost be enough to make meaningful changes?

    The lack of finalized standards for the smart grid is another piece that comes into play. No utility is terribly willing to make drastic changes to their systems with the fear that they may have to change them again once DOE standards are set in the future.

    Another issue is whether or not the consumer is ready to see this kind of rapidly-changing data. The swings from base load to peak power can be great and are constantly changing with the market. If this information is fed too quickly, customers with good intentions of being conscientious of their energy usage may experience information overload. The result is discouragement and cessation of use of these great tools.

    • Would anyone be interested in a home application device that does this for you? a by the minute representation in dollar amount of how much energy your home is consuming?

  3. Could we start with smart transformers that don’t explode on a regular basis? Or smart distribution lines, that don’t cause outages during moderate weather? It’s not like the Bay Area has tornados or ice storms.

    As to real-time data, what’s so difficult? If you are interested in real-time data, you most likely already have Internet access, the utility doesn’t need to provide it. A wireless webcam pointed at the meter and some OCR software would provider real-time data for under $50 per meter.

  4. Steven Schoch

    I have a TOU meter (that came with the house in 1996). It’s rather old, and not smart. All my neighbors’ meters (in Sunnyvale, CA) have been replaced with Smart Meters, but mine has not. Is PG&E having trouble moving existing TOU customers to the new Smart Meters?

  5. Dick Zeren

    And of course, there is no tariff for true, real-time pricing, as a Commissioner pointed out at a conference late last year. Getting any new tariff is a contentious and time consuming process. With various classes of customers, “consumer” advocates, merchant energy providers, the ISO, the utilities (including those not regulated by the CPUC), and who knows who else arguing, it could take years.
    The CPUC workshop is just an early step in the process.

  6. At first blush, Krevat’s claim that you would need broadband in order to do near realtime usage or pricing information seems far-fetched. We’re talking about an internet packet or two every few minutes — even dial-up speeds would be extreme overkill.

    • Jon Husen

      The current figures of truly real-time data transmitted for 1.4 million smartmeters is 100 petabytes annually. To that you have to add data from the utilities to the meters for pricing information and demand-response instructions. Utilities may also have real-time usage data accessible via a web portal or smartphone app. There is far more information being transmitted than initially meets the eye.

  7. Louis Alten

    Good article.

    One correction: the existing smart grid systems are not “designed to be read every 15 minutes or hourly”. The power usage is logged in 15-60 minute intervals and recorded on the meter. Meters communicate this usage information much less frequently to the utility, at most every 4 hours.