There's a lot of stats and utilizing personalized information that go into this....

There's a lot of stats and utilizing personalized information that go into this.

I was part of https://www.oracle.com/industries/utilities/products/opower-... before the oracle acquisition, and we powered the Demand Response program, in addition to other things.

There's a couple steps to this:

    1. Eligibility and Program Opt-outs (figure out who it is legal to send the message to)
    2. Machine Learning to predict the ROI - how much to offer and how many people are likely to "advertise" to - the utility needs to know so they can keep within budget.
    3. What effective, A/B tested messaging to send (some people care about money back, some cared about being green, some both)
    4. What medium the customer reacted to (we did batch, usps snail mail - generate PDFs and automated the mailing while minimizing postage cost, or did they react to SMS)
    5. The actual platform to blast out texts to all residential customers 24hrs before the DR (Demand Response) event to a city - ex: (BGE) boston, tokyo, PGE (bay area)..

fun times.