Expand my Community achievements bar.

SOLVED

Positive vs Negative inclusion within segments

Avatar

Level 2

Good morning,

I have a technical question around best practice with building segments.

I have an offline trait from my data warehouse, for customer’s product holding and 3 other traits for lifecycle.

I want to segment the customers with product ‘A’, into 4 product lifecycles. Early life, in life, recontracting window and out of contract.

For early life, recontracting window and out of contract I have made positive inclusions, however for in life I have removed all the other options i.e. ‘AND NOT’ for early life, recontracting window and out of contract. (Negative inclusion)

My assumption is that Product A has 10m devices, if I remove 3 options, say totalling 4m, then I’ll be left with 6m. This does not seem to be the case as I’m seeing good volume in all other segments apart from in life.

The acid question is – do you always need to make positive inclusions rather than assumed inclusion by removing all other options? (i.e. does it handle remainders?)

Many thanks,

Pete

1 Accepted Solution

Avatar

Correct answer by
Employee

Hi Pete,

Is your definition like the below?

Has Product A

AND

(Customer is scored between 80 - 90 OR Customer is scored between 90 - 100)

AND NOT

(CONTRACT DATES early life 0 - 30 OR CONTRACT DATES early life 30 - 60 OR CONTRACT DATES recontract window 0 - 30 OR CONTRACT DATES recontract window 30 - 60 OR CONTRACT DATES out of contract 0 - 90)

if not the AND NOT will be applied to "CONTRACT DATES early life 0 - 30" only.

Also you can check the overlap between the traits from UI Analytics > Audience Reports >Overlap Reports> Trait to Trait overlap Report.

Thank You,

Shubha

View solution in original post

4 Replies

Avatar

Employee

Hi Pete,

Have you tried looking at overlap between these traits. Also can you let me know what conditions are you using in between the exclusion trait rule?

Thanks,

Shubha

Avatar

Level 2

Hi Shubha,

Thank you for taking the time to reply.

What would be the best method of checking for overlaps in this instance?

The segment looks like this:

Has Product A

AND

Customer is scored between 80 - 90

Customer is scored between 90 - 100

AND NOT

CONTRACT DATES early life 0 - 30

CONTRACT DATES early life 30 - 60

CONTRACT DATES recontract window 0 - 30

CONTRACT DATES recontract window 30 - 60

CONTRACT DATES out of contract 0 - 90

All of the contract dates are exclusions, which should leave the remaining few customers in the pot as a remainder.

Having created another segment that includes a set of product dates as a positive inclusion i.e:

Has Product A

AND

Customer is scored between 80 - 90

Customer is scored between 90 - 100

AND

CONTRACT DATES inlife

AND NOT

CONTRACT DATES early life 0 - 30

CONTRACT DATES early life 30 - 60

CONTRACT DATES recontract window 0 - 30

CONTRACT DATES recontract window 30 - 60

CONTRACT DATES out of contract 0 - 90

I find that the volume is much higher on the segment with the positive include vs the negative inclusion. Volume being RT vol and total vol.

(Note i created both segments at the same time).

Many thanks for your help.

Cheers,

Pete

Avatar

Correct answer by
Employee

Hi Pete,

Is your definition like the below?

Has Product A

AND

(Customer is scored between 80 - 90 OR Customer is scored between 90 - 100)

AND NOT

(CONTRACT DATES early life 0 - 30 OR CONTRACT DATES early life 30 - 60 OR CONTRACT DATES recontract window 0 - 30 OR CONTRACT DATES recontract window 30 - 60 OR CONTRACT DATES out of contract 0 - 90)

if not the AND NOT will be applied to "CONTRACT DATES early life 0 - 30" only.

Also you can check the overlap between the traits from UI Analytics > Audience Reports >Overlap Reports> Trait to Trait overlap Report.

Thank You,

Shubha

Avatar

Level 2

Hi Shubha,

That's correct, it looks like how you've defined it above.

Just for my understanding, "if not, the AND NOT will be applied to "CONTRACT DATES early life 0 - 30" only."

What would this segment look like? And why would it only look at the first statement?

Cheers,

Pete