3 ms·

This is the opposite of AI use cases in marketing. You are declaring a specific timeframe for your message delivery. That is not how a marketer should deploy AI

This is the opposite of AI use cases in marketing. You are declaring a specific timeframe for your message delivery. That is not how a marketer should deploy AI. I haven’t been in any pitch meetings since AI assclownery took hold so I can’t comment on how the term is being abused. What I can say is that a model that used AI would take every parameter it could about each customer and determine the optimal time to sent an email to get a conversion. The only inputs the marketer should provide is raw historical data with clear parameters like order value, order items, estimated revenue, buyer classification, a stream of subsequent etc and date, and the model should solve for the correct timestamp to send the follow up message. I don’t think the AI is writing the message yet, and I don’t think you need a neural net to do a decent job at solving for the right datetime to send. I do think the approach I described would get superior conversion rates than a rule, cost more to make than that rule, and definitely demand a decently huge dataset to add much value.

taeric 9y ago
I think the big problem with your example is that most companies just don't have something I want to talk with them about. For large stretches of the year. No amount of ai is going to change that I am not in the market for a car, as an example. Even biking gear is off limited sellability to me most years. Hard to know when I'm going to buy new booking shoes, unless you know when my shoes are going to go bad.
zer00eyz 9y ago
Though your generally right, the article sets a tone in the first paragraph. It is saying that if your looking for ML/AI solutions for marketing and you ARENT doing the basics already then your throwing good money after bad. You should START with some sql and targeted emails before you dive into a large and potentially expensive project.
nostrademons 9y ago
I think the main point of the article is that 99% of the value in that is in a.) sending the follow-up email at all (which just takes a cronjob) and b.) identifying which customers to send that follow-up e-mail to (which just takes a SQL query). While it's probably nice to try and predict the ideal time to send it, the gains you get from that are marginal compared to steps a & b, which many companies aren't even doing today.

taeric 9y ago

I think the big problem with your example is that most companies just don't have something I want to talk with them about. For large stretches of the year. No amount of ai is going to change that I am not in the market for a car, as an example. Even biking gear is off limited sellability to me most years. Hard to know when I'm going to buy new booking shoes, unless you know when my shoes are going to go bad.

zer00eyz 9y ago

Though your generally right, the article sets a tone in the first paragraph. It is saying that if your looking for ML/AI solutions for marketing and you ARENT doing the basics already then your throwing good money after bad. You should START with some sql and targeted emails before you dive into a large and potentially expensive project.

nostrademons 9y ago

I think the main point of the article is that 99% of the value in that is in a.) sending the follow-up email at all (which just takes a cronjob) and b.) identifying which customers to send that follow-up e-mail to (which just takes a SQL query). While it's probably nice to try and predict the ideal time to send it, the gains you get from that are marginal compared to steps a & b, which many companies aren't even doing today.