3 Replies Latest reply on Feb 28, 2017 9:11 PM by Sanford Whiteman

    Filters for data feed updates

    Erin Mackintire

      Hi Everyone,


      I am looking for suggestions in this scenario where we can't seem to find the easiest solution for this. We have an Informatica feed that updates data daily and daily batch campaigns to send emails based on the qualifying data. If the informatica feed fails that day and the data is not updated correctly, we want a safety net to make sure that these campaigns are not triggered as they could be sent to the wrong audience. I thought of usind "data value changed" filter, however, in some cases the data will be updated for some records and have failed for others. We also do not want to create a new dummy field since this could also cause issues with the informatica field.

      Does anyone have a suggestion on what we could build out to make sure that these campaigns run smoothly?

      Thanks for your help!

        • Re: Filters for data feed updates
          Joey Forcelli

          Hey Erin,


          So informatica updates fields on the leads and then the batch campaigns are run based upon that updated information, correct?  Why can't you use the data value changed filter?  You could setup the smartlist to only include leads where the data value has changed within the last 23 hours.

            • Re: Filters for data feed updates
              Erin Mackintire

              Hi Joseph,

              Thanks for responding. I think that would be my last ditch attempt, but I would rather find something to build that was all or nothing. We're usually updating thousands of records and a bunch of fields at once so there isn't one field I could choose as the "data value was changed" field. And we are running 4 jobs, so some of the fields on a record may be updated and some may not.

                • Re: Filters for data feed updates
                  Sanford Whiteman

                  Erin, there's a certain impracticality to trying to detect whether a multi-row import job, let alone a set of jobs, was successful as a unit since Marketo doesn't consider a "missing" update to be, well, missing.


                  What you should do instead is schedule the campaign on the client app side, only when the app considers the job(s) done. Only that side (or a separate monitoring app that's tracking that app) can judge completion.



                  P.S. This is a common problem across systems and databases, not just w/Marketo. If an essentially passive server (like a SQL database) doesn't get any inbound connections for a period of time, it doesn't have any reason to know that anything's missing. Only the client app can know if an inbound job failed to start or complete.