| Poll |
Does real-time data improve predictions of consumer behavior, in your experience: [77 votes total]
|
|
Significantly improves (9) |
12% | |
Somewhat improves (21) |
27% | |
No effect (7) |
9% | |
No experience with real-time data (40) |
52% | |
Comments
Greg James, Real-Time Data
I have been using bank account transaction-level data for over 4 years.
I process it daily, so that does not quite constitute real-time in the
strictest sense. My experience has shown three things: understanding
behavior at this level is an order of magnitide more complex than
summary data; maturity using behavioral data can be observed as one
moves from reactionary to contemporary to anticipatory models and
actions; and exploiting this information requires systems and processes
that are highly integrated and can all evolve simultaneously over time.
Gregory Piatetsky-Shapiro, Using Real-time Data for Prediction
Predicting consumer behavior for CRM and direct marketing (what the
customer likely to buy, is she likely to attrite, etc) was traditionally
done using monthly data. Recently in e-commerce and other fields
companies are using real-time data (last click, latest call or
transaction) to generate a real-time prediction.
What was your experience with real-time data for prediction, if any ?
| |
|