If we see a consistent increase in volume (i.e. 10% more offered) week over week for several weeks, will Automatic Best Forecasting be smart enough to anticipate approximately 10% increase week over week moving forward?  Or will it try to average it out? 

Hi Shelby,

I am reviewing unanswered posts, did you get an answer to this question?

If we see a consistent increase in volume (i.e. 10% more offered) week over week for several weeks, will Automatic Best Forecasting be smart enough to anticipate approximately 10% increase week over week moving forward?  Or will it try to average it out? 

Hi Shelby,

I am reviewing unanswered posts, did you get an answer to this question?

If we see a consistent increase in volume (i.e. 10% more offered) week over week for several weeks, will Automatic Best Forecasting be smart enough to anticipate approximately 10% increase week over week moving forward?  Or will it try to average it out? 

Hope you dont mind me jumping on your thread Shelby, but as a new user to Genesys Cloud, I am also interested to learn more about how automatic best method will treat situations like this. Also wanting to know more about how it treats outlier data. I am used to over many years being able to enter overlays (in pure engage) to ensure forecasting disregards outliers where things such as system outages/external impacts are driving volumes significantly higher or lower than 'normal'.

Hi Shelby,

I am reviewing unanswered posts, did you get an answer to this question?

If we see a consistent increase in volume (i.e. 10% more offered) week over week for several weeks, will Automatic Best Forecasting be smart enough to anticipate approximately 10% increase week over week moving forward?  Or will it try to average it out? 

Hi Tristan, 

Another great question! We handle outliers via the outlier and calendar effect detection and pattern detection phases of creating an ABM forecast. We look for three specific outliers: random pulses, temporal changes and level shifts. We then look to see if the outlier re-occurs during a particular time of the year, month, week, day and so on to identify if it is genuinely an outlier or a trend. 

We understand that ML/AI can only do so much, so we do have it on our roadmap to give users the ability to help the ABM by tagging historical outliers and telling ABM how it should be handled for future forecasts. Unfortunately, we do not have a GA date for this feature as it is in the very early stages, but it is on our current roadmap. Could I contact you for any research purposes once we start actively working on this feature? 

Thanks,

Hope you dont mind me jumping on your thread Shelby, but as a new user to Genesys Cloud, I am also interested to learn more about how automatic best method will treat situations like this. Also wanting to know more about how it treats outlier data. I am used to over many years being able to enter overlays (in pure engage) to ensure forecasting disregards outliers where things such as system outages/external impacts are driving volumes significantly higher or lower than 'normal'.

Hi Shelby,

I am reviewing unanswered posts, did you get an answer to this question?

If we see a consistent increase in volume (i.e. 10% more offered) week over week for several weeks, will Automatic Best Forecasting be smart enough to anticipate approximately 10% increase week over week moving forward?  Or will it try to average it out? 

Hi Tristan, 

Another great question! We handle outliers via the outlier and calendar effect detection and pattern detection phases of creating an ABM forecast. We look for three specific outliers: random pulses, temporal changes and level shifts. We then look to see if the outlier re-occurs during a particular time of the year, month, week, day and so on to identify if it is genuinely an outlier or a trend. 

We understand that ML/AI can only do so much, so we do have it on our roadmap to give users the ability to help the ABM by tagging historical outliers and telling ABM how it should be handled for future forecasts. Unfortunately, we do not have a GA date for this feature as it is in the very early stages, but it is on our current roadmap. Could I contact you for any research purposes once we start actively working on this feature? 

Thanks,

Hope you dont mind me jumping on your thread Shelby, but as a new user to Genesys Cloud, I am also interested to learn more about how automatic best method will treat situations like this. Also wanting to know more about how it treats outlier data. I am used to over many years being able to enter overlays (in pure engage) to ensure forecasting disregards outliers where things such as system outages/external impacts are driving volumes significantly higher or lower than 'normal'.

Hi Shelby,

I am reviewing unanswered posts, did you get an answer to this question?

If we see a consistent increase in volume (i.e. 10% more offered) week over week for several weeks, will Automatic Best Forecasting be smart enough to anticipate approximately 10% increase week over week moving forward?  Or will it try to average it out? 

Hi, @Daniel Chapdelaine!  

Could I possibly volunteer to assist with testing and research for this roadmap item?  We're having the hardest time getting our forecast under control due to process improvements we are making to our business; some weeks we have 50+ manual modifications on the forecast.  I shared this thread with my director, and he seems interested.

Hi Tristan, 

Another great question! We handle outliers via the outlier and calendar effect detection and pattern detection phases of creating an ABM forecast. We look for three specific outliers: random pulses, temporal changes and level shifts. We then look to see if the outlier re-occurs during a particular time of the year, month, week, day and so on to identify if it is genuinely an outlier or a trend. 

We understand that ML/AI can only do so much, so we do have it on our roadmap to give users the ability to help the ABM by tagging historical outliers and telling ABM how it should be handled for future forecasts. Unfortunately, we do not have a GA date for this feature as it is in the very early stages, but it is on our current roadmap. Could I contact you for any research purposes once we start actively working on this feature? 

Thanks,

Hope you dont mind me jumping on your thread Shelby, but as a new user to Genesys Cloud, I am also interested to learn more about how automatic best method will treat situations like this. Also wanting to know more about how it treats outlier data. I am used to over many years being able to enter overlays (in pure engage) to ensure forecasting disregards outliers where things such as system outages/external impacts are driving volumes significantly higher or lower than 'normal'.

Hi Shelby,

I am reviewing unanswered posts, did you get an answer to this question?

If we see a consistent increase in volume (i.e. 10% more offered) week over week for several weeks, will Automatic Best Forecasting be smart enough to anticipate approximately 10% increase week over week moving forward?  Or will it try to average it out? 

If we see a consistent increase in volume (i.e. 10% more offered) week over week for several weeks, will Automatic Best Forecasting be smart enough to anticipate approximately 10% increase week over week moving forward?  Or will it try to average it out?