Question

Rules Horizon Feedback

  • 15 February 2023
  • 18 replies
  • 195 views

Userlevel 5
Badge +9

Anyone else started using Rules Horizon?  I created my first one today and overall it’s pretty slick.  I love to Update CTA feature.  We do have some feedback we are gathering, though, and I wonder if it is best to post here, the general Community, or both?


18 replies

Userlevel 7
Badge +11

Started using it, definitely has some strong improvements. I also feel like there are some additional clicks in certain spots that I am not quite able to pinpoint yet or even validate. Execution History for one. Need to get used to it...

Userlevel 4
Badge +5

Just tried yesterday for a S3 import. For now looks fine, but still need to investigate further.

Userlevel 6
Badge +9

You are allowed to say my mantra is cloning everything because it’s true, but I would like to be able to clone a branch instead of cloning an action.

It doesn’t make sense to build the filter, then the action and being able to add another action from that filter and then having to filter differently, and then a similar action.

So I want to clone branches so I can just edit the filter without having to rebuild the action. 

 

 

Userlevel 7
Badge +6

@anirbandutta where are we with this? 

Opening up Beta Feature Conversations Public

This has come up a few times in the community especially in the context of repeating feedback from closed beta groups. Would at least be nice to know what direction it is going, if anywhere.

Also, Snowflake as a source for us is $$$.

 

@angela_domenichelli I really like the Data Designer UI way to visualize the rule design and the ability to preview merges and transformations. 

 

Like @gunjanm said though there are some pain points (most of which I’ve brought up in the beta group) among them are:

  • Two ways to build rules now (within Horizon Rules) are clunky, probably confusing for new admins and not really a value add IMO.
  • Way more clicks to do pretty much anything.
  • The way that rule filters have been separated from rule actions is an operationalized disaster (from a post I made in the Beta Group)

    You also, partially as a result, can’t clone a rule action unless that action is part of the same filter action branch. Cloning an action really only works if it’s filtered the same way. 

I’m also concerned at the timing of how soon Bionic Rules will be shut off for making new or cloned rules given the lack of maturity of HR at this point and any sort of clear migration process. 

 

Overall I think it’s going in a great direction, I just would hate to see such a core and integral feature of Gainsight be rushed at the expense of usability, functionality and admin quality of life.

Userlevel 6
Badge +3

@rakesh FYI 

Userlevel 7
Badge +2

@bradley,the 1st question I share my summary

@anirbandutta where are we with this? 

Opening up Beta Feature Conversations Public

This has come up a few times in the community especially in the context of repeating feedback from closed beta groups. Would at least be nice to know what direction it is going, if anywhere.

 

For your second question

Also, Snowflake as a source for us is $$$.

Do we have an Idea for it? I can try get some answers.

Userlevel 6
Badge +9

Further to my cloning post yesterday and feedback re. how I would find it easier for filters to continue to be joined with actions (or at least to be able to clone a “filter + action” branch) and to @bradley’s comment, I have different feedback re. new admins. 

When I built my first horizon rule, I thought, “wow, this is more guided and it is slightly annoying now that I have developed my own habits, but for a new admin, this extra guidance would make it easier apprehending rules” (because I recall being a bit scared when I first started with rules 18 months ago and I would have appreciated this “being on rails and not being able to derail experience”). So in short, I actually think the new RE would facilitate a new admin’s life. 

Userlevel 7
Badge +6

 

For your second question

Also, Snowflake as a source for us is $$$.

Do we have an Idea for it? I can try get some answers.

That wasn’t a question 🙂 I was saying Snowflake as a data source for rules is awesome ($$$ = money = awesome)

Userlevel 6
Badge +9

@anirbandutta Following on @bradley ‘s ask, I’ve seen that more people on Slack are looking for this post. Can we move it outside of this group or should a new one be started for everyone to partake? 

Userlevel 7
Badge +2

Ah…  thanks Alizee and Bradley for mansplaining to me the ask.

For this thread, allow me a day or two to check on alignment and then we should have no problems to make things public on a thread by thread basis.

Userlevel 7
Badge +2

Ah…  thanks Alizee and Bradley for mansplaining to me the ask.

I mean I missed the part where I could have acted and I appreciate you underlining it for me.

P.S. I’m already checking about opening up this particular feedback.

Userlevel 7
Badge +2

Thanks Product Council Admins for starting it off

@rakesh for reviewing

@bradley and @alizee for prompting to include the wider community - This Conversation is opened up for all the Rules-users to discuss.

P.S. with threads being opened from Private: Some of the comments may be indicative of a past state of the product and may be of historic significance.

Userlevel 5
Badge +5

Agree with all of the above:
Pros:

  • Data designer UI to build the dataset is awesome!
  • Because of this, the formula fields are "easier" to find (A-Z order at the general level, vs A-Z order by datatype in Rules Engine)
  • And also, the "to string" fx that was only available in DD, now can be used in a rule

Cons:

  • The Select an Object or Prepare Dataset action is weird/annoying. If you chose the former and then realize you need something else, you have to start from zero again.  So basically even using just one source I'll always choose the Prepare Dataset option just in case.
  • The lack of capability to clone a branch
  • The filter before the action is at the very least, inconvenient, any change in the filter will mean rebuilding the action


I just started building some simple rules so my insights are very limited, but I’ll circle back if I have any other feedback to share.

Userlevel 4
Badge +1

Overall, I like the new UI and being able to preview data sets with transforms/formulas without having to run test runs. 

Echoing the paint points that other mentioned: the amount of extra clicks, inability to clone branches, and the process for setting up criteria. Some of our bionic rules have ~20 actions with unique criteria. I’m not looking forward to manually creating a branch for each and every action 😅

Userlevel 6
Badge +9

Additional feedback on date filtering parity which I believe must have been shared elsewhere: date filtering in horizon rules was much less annoying than DD, combination of filters to achieve a Subtract or Add N Days from rule date is hellish. 

 

Userlevel 2
Badge +2

PROS

Overall, I really like the new interface! It’s much more user-friendly and I am a very visual person, so find the below items tremendously impactful.

  • Data Design functionality is AMAZING! It’s a much cleaner experience and I love the consistent interface between the features
  • Being able to visualize which dataset each action is running off of is also super helpful - especially when diagnosing failures. BIG fan of this!
  • The execution logs are also very intuitive 

CONS:

I have already opened several support tickets for which I was told that the issue is a bug that Gainsight cannot fix. These bug fixes only fix net new datasets or net new rules, but not existing datasets/rules. The issues I’ve experienced so far:

  • For extracts to S3, Display Name values are not appearing as column headers, regardless of being set to different values than Field Names
  • One rule I built kept failing with the error: “Improper reference of taskId found in action.” This rule had a dataset that we deleted and then rebuilt the design within the same rule. The only solution was to delete the rule and build it again. GS did confirm a bug fix was shipped for this on March 10th and new rules should not experience this issue
  • I keep experiencing strange behavior when using the “Duplicate String Value” option for dropdown fields. I’ll select that option, and save the rule. Then, I’ll delete the original dropdown field & change the display name of the string field to the field name of the dropdown field, and Save again. This appears to work, but then randomly the dropdown fields will reappear & the string fields will disappear. I have at least 3 rules where I’ve replicated this behavior successfully. 

Overall, I think the redesign is fantastic, and I’m looking forward to all of these strange bugs getting fixed so that my rules perform as expected & I don’t have to keep rebuilding them. 😅

Badge

Hey all,

I really like the Data Designer UI, it’s pretty handy to create rules.

I’m not that happy with the Action handling that was implemented, if you have multiple actions in different dataset tasks it’s not obvious which one is excecuted first. If you agree please vote here: 

 

 

Another small point I don’t understand is, why Picklist IDs are prioritized over Picklist Values. The ID’s of a picklist item is never used and not that easy to find. The Picklist Values as a direct string would be the way to go.

 

We’ll try using only Horizon Rules from now on. I hope that a few teething problems will be resolved before ending the BETA.


Best

Christian

Hi All,

 

Some feedback this end too - specifically around Horizon Rules and Relationships as well as some general UX thoughts.

 

From an admin perspective, I see the improvements and like the ability to preview as well as aligning the experience between Rules Engine and Data Designer, however, I have found a couple of pain points:

  • Mandatory Lookups for Load to relationship etc. are slightly less easy to do - previously these were separated out into their own section nested under the action, however now it feels like it’s ‘map the field and hope’. I can imagine that for newer admins it’s confusing/stressful when trying troubleshoot rule failures, that wouldn’t fail on a test-run.
  • I feel like I’m having to click through more menus/action points are nested deeper than previously - some of this will be bias/muscle memory but I find there are more steps required here - eg filtering the action separately from the action itself.

 

I will say, however that I find the ability to just fire a rule with from a general object vs. picking fields quite useful when I don’t need ultra-specific actions/joins and am just looking to change the defaults of some Booleans for example.

 

Best,

 

Leigh

Reply