Mojo, Sales Dialer or Stellar Prospects?

[Josh]

First, are you joking about filing a patent for a system that scales the lines dynamically? I think someone already beat you to the punch on that one, in fact, if I understand what you're talking about correctly, it's a trick my hardware already does

That said, I do agree with you about burning through data, except for the point when you have records you've just been calling and calling and calling. I do think it's bad business to have a 3% or higher drop rate (last night mine was around 2.2% and I didn't like it, but it was a necessary evil.)

Just a few other points:
-5-7 telemarketer is where predictive starts to make more sense.
-Going as high as 8 lines per caller can make sense if you're using really horrible data (Yes, this can be done and have a lower than 3% drop rate).
-The FCC wouldn't show up on a small shop, especially if it's in the Philippines.
-If the FCC shows up technically you are supposed to provide them with all of your call logs, but it's easy enough to heal up your drop rate by just keeping an eye on what's going on.
-If the FCC shows up and you had some type of hardware failure that wiped out your logs, that sounds like a perfectly reasonable explanation for what happened.

 

[ksigmtsu]

You'll have to figure out how to scale it without interacting with databases.

I know a method, it's attached to your hand. The automation on it sucks.

---

It's pretty specific, I think the document is about 12 pages worth of explanation, and there wasn't a product on the market that did exactly what I'm saying already documented anywhere in the world.

Also a patent wouldn't prevent you from doing it yourself, however it would prevent other people from selling it without a royalty.

 

[djs]

Actually, dynamic line scaling has been around for a LONG time. Vicidial calls it 'taper'. Every true predictive dialer has it.

What you use to 'predict' for the dynamic scaling can be unique and is where I thought you were going with your patent. This is why predictive dialers don't work for small call groups (i.e., single agent, or even up to 4-5 agents), to much stuff can happen on a call that will cause the 'prediction' to fail.

Dan

 

[Josh]

You'll have to figure out how to scale it without interacting with databases.

I know a method, it's attached to your hand. The automation on it sucks.

So you're saying call A leads on a 2:1 ratio, B leads on a 3:1, or that's the gist of it? Or are you saying to have it adapt relative to the drop rate with the same type of list?

 

[ksigmtsu]

So you're saying call A leads on a 2:1 ratio, B leads on a 3:1, or that's the gist of it? Or are you saying to have it adapt relative to the drop rate with the same type of list?

Its adapting related to the connect ratio of the list data in the specific campaign, number of call attempts, age, number of drops overall, and position in sales process. It can also mix lists, and prioritize internally in the same campaign and dynamically scale while doing that.
- - - - - - - - - - - - - - - - - -
I had some AI classes in college.
- - - - - - - - - - - - - - - - - -

Actually, dynamic line scaling has been around for a LONG time. Vicidial calls it 'taper'. Every true predictive dialer has it.

What you use to 'predict' for the dynamic scaling can be unique and is where I thought you were going with your patent. This is why predictive dialers don't work for small call groups (i.e., single agent, or even up to 4-5 agents), to much stuff can happen on a call that will cause the 'prediction' to fail.

Dan

Yeah its the data being used to determine the scaling of the lines and priority that didn't exist. It was a static or taper, with a scale down based on drop % ratio or triggered event, but there was not any algorithmic method that focused on single agent connect rate and reduction of drops on certain data to 0 yet scaled up lines when calling other types of data in relation to the existing data related to prior experience.

 

[djs]

Got it, but then there is no problem with using 6 lines (other than common sense) and scaling based on drop ratios. This won't run afoul of your patent pending. It is in fact, what predictive dialers do.

Dan

 

[ksigmtsu]

In theory the system could choose to scale up to 100 lines if I didn't cap it at 4, in order for it to generate a specific % chance of an answer as long as the drop % stayed below 3%. The more data I have the more I can relax that restriction.

 

[djs]

yes, but there are practical limits.... 3 lines is good for me. Maybe 4. I don't know that it is productive to get above 4 at a time. Of course, that is what you said way back in post 9 of this thread....

Dan

 

[ksigmtsu]

My math starts showing that it does not increase your answer chance more than it increases drop chance at a certain point. Just below that is the sweet spot but it is a moving target.
- - - - - - - - - - - - - - - - - -
Can be very different if you do a press 1 campaign or something like that also.