What is our primary use case?
Developing and operationally implementing a powerful lead scoring model for a major Multiufamily developer and operator of apartment properties throughout major western states. The work included 3 years of data across over 60 properties with more than 500,000 leads and 3 million transactions.
How has it helped my organization?
Increased sales force productivity by permitting them to prioritize activity during peak leasing periods on those leads most likely to close
What is most valuable?
The ability to do the templating and be able to transfer it so that I can easily do multiple types of models and data mining is a valuable aspect of this solution. You only have to set up the flows, the templates, and the data once and then you can make modifications and test different segmentations throughout.
We were working across a number of internal departments as well as some outside departments and this solution made it extremely easy to communicate across functional area because it was all in flow chart and data form so that if somebody had an issue, like changing the data set or something like that, they could point right to it and we could get that handled and incorporated into the model. It's extremely efficient on the computer. We had to do a number of resets on the data in the model and to be able to turn things around and validate the model and the new set in two hours, was just incredible for me.
It was very robust. The ability to move the objects around so easily and then communicate is really its power. Then to be able to show it to the sales and senior management, in terms of what was employed and made it very easy to get my job done.
What needs improvement?
In terms of improvement, I'd like to have more ability to understand the detailed impact of the variables on the model and their interactions. Their algorithms are very powerful and they explain overall the net contribution of each of the variables to the solution. In terms of being able to say to people "If you did this, you'll get this much more improvement" Azure (at least my understanding of it) doesn't provide readily accessible tools to assess from a management perspective the impact of their changing a sinimized, the better.gle value - for instance in closing a lead, decreasing response time by 10%.
I recognize that the multivariate algorithms used from decision trees to neural nets do not readily provide the coefficients for each variable ala the older regression modeling approaches. My experience over my 50 years of developing and implementing predictive models has been that more than half the value of modeling lies in improving management's understanding of the process being modeled, often leading to major organization and operational structure changes. More ability to understand the variables impacting the end result being optimized would be very useful.
For how long have I used the solution?
I have worked extensively with this solution for the last three years.
What do I think about the stability of the solution?
I haven't had any problems with stability.
What do I think about the scalability of the solution?
I didn't have any issues with the scale. we rapidly went from test to full implementation across all datasets.
How are customer service and technical support?
I never had to use technical support.
Which solution did I use previously and why did I switch?
I have used SPSS modeler (part of WATSON really) but because client was a Microsoft shop, I switched to Azure.
How was the initial setup?
I found the setup to be very easy. I've been doing this type of work for 50 years so the modern terminology isn't always the same as what I grew up with. It took me a while to understand that, but the setups were very easy. As with anything, the hardest part is always getting the data together, but the outside consultants had built up a very, very good data warehouse. The ability to manipulate the data and create variables was very nice.
THIS IS THE ONLY MODELING APPROACH THAT EVER WORKED THE VERY TIME I RAN IT!!
What's my experience with pricing, setup cost, and licensing?
Because client isa Microsoft shop, everything was Microsoft in terms of having solutions like Power BI and stuff like that. Azure is very useful and very inexpensive.
What other advice do I have?
The major advice I give is that clients must get the user,somebody who understands the business issues, to be deeply involved with it and the data transformation. Most people don't. And that's true for data science applications. We don't just follow the data in a big pile and remodel, we advance the process that we're modeling. Consider what transformations of the data you need to make it workable and usable.
Remember, over half the initial value of modeling is the strategic understanding provided re the importance of different variables to the model and hence the organizaion's performance. Very often the modeling identifies opportunities for changing structures, decision rules, etc. even prior to the model's actual implementation technically.
I would rate it a nine out of ten.
Which deployment model are you using for this solution?
If public cloud, private cloud, or hybrid cloud, which cloud provider do you use?