I host http://www.nodewatch.co.nz/, a website for electricity users taking spot pricing, or small generators taking spot pricing that sends alerts to those users when prices pass through thresholds the users can set.
The signed-up users appear to be a mix of Flick customers, industrial users and small generators, all of whom wish to react rapidly when the price goes high, either by turning things off, or by turning generation up.
Recently, there have been a number of short-duration prices spikes, like the one on the second of June. This has caused two things: one, a flurry of sign-ups to nodewatch, and two, questions about the timeliness of the data. If a price spike lasts a single trading period, it's very important that the alerts get to the users rapidly.
(For example, see this discussion on Geekzone.)
Currently, I scrape five-minute prices from WITS free to air (and am *fully* aware of the difference between real-time prices and final prices, we don't need to resolve that here). There are a number of issues with this (scraping is awkward and slow), but the current issue is timeliness of the data: by the time I get the price from WITS about half a trading period has passed. I understand the EMI real-time data APIs are similarly delayed.
As I said, I understand real-time prices are not the same as final prices, but at this time, they are the best thing nodewatch users have to go on. Large generators in the industry have *plenty* of tools to help manage (and forecast) these prices. It would be really helpful to have timely data on real time prices available for the small players and customers to enable them to respond before it's too late.