Einstein Analytics Camp Fire

When I’ve been to London last time, for the CloudCraze training, I found time during evenings and did my Einstein Analytics Advanced Acreditation. And thanks to this „investment“ I had privilige to spend two days in the Salesforce Tower in London last week to attend the Einstein Analytics Camp Fire. These were two super intensive days, as Ziad packed into them the content of maybe 5 days of regular training. Super intensive and super informative, I don’t remember when I got such overwhelming amount of information in such short time before. A lot of tips included on top of that.

How to drive adoption – CASES

When you want to drive adoption, it is pretty easy, just follow some rules. Support collaboration on dashboards (teach users how to annotate and share), make them actionable (don’t forget to enable actions), self-service (yeah, this is about training as well), embedded it in Salesforce/Community (I can hear your screaming – extra licences) and smart design (which is super important as I remember from Keboola Hackathon). If you don’t have enought time to read this article find a minute to watch this video.

I asked how much time they think it took and was surprised to hear that it is just a few hours, day a max. And it looks super cool, I would love to use it every time. I mean several times a day.

Data layer

Enable replication (in Salesforce setup) but don’t forget, that incremental replication doesn’t delete data. Don’t create one huge data set but rather more smaller and use as many fields as possible from parent objects.

Use recipes heavily, you can use bucketing for setting order of values (if you have just a few of them, doesn’t make sense for hundreds), use it for combining more data sets, for dates to avoid static steps and bucketing (wonder what I meant with this note, but it was important).

Data flow can do all the heavy lifting, so you don’t have to do much with the data later on. ComputeExpressions and computeRelatives (and all the other functions) also means you can do a lot of stuff on your own and don’t have to ask Salesforce admin to update/change the data structure. And you can also use it for deduplication or snapshotting 🙂

In data flow put filter as separate node, not on sfdcDigest node as it will affect all sfdcDigest of the same object. Don’t let everyone edit data flow, just a small group of people let do it, who know what they do.

You can always „connect“ your datasets via connections between widgets, don’t need to combine them all into one big matrix.

You should know about Dataset Utils as it can save you some time.

And when I saw the data flow editor I wished Salesforce dev teams cooperate more closely or maybe create some frameworks internally. Journey builder, Engagement Studio, Lightning Flow and this all do „about the same“ and when I saw the plans for Lightning Flow I’ve been really excited. The ability to chat on every elements, group them together, so they don’t so much space on screen and much more would be awesome to have it those other editors as well.


  • maximize the number of columns, so they are more narrow and you can insert small images;
  • hide left axis title when not needed;
  • in lens you can add measure * and rename it to the name of the object, so it shows Count of ObjectName instead of Count of Rows;
  • don’t forget to name your steps, so you know which is which;
  • do most of the customization on widgets not steps (reference lines for example);
  • when you edit column in lenses it has some super features such as ranking, running total, sliding window and other functions;
  • you can use marker for notes, it looks great and it is more visible than annotations;
  • create hidden columns for sorting;
  • maps are super cool, use them and create a custom one;
  • use pages, copy elements between them and create links, it adds a lot of dynamics and beauty to your dashboards.

Install apps to learn new things

Einstein Analytics Sales Analytics template and Einstein Analytics Learning Adventure app and Einstein Analytics Learning Map (primarily for clients) are really great to show you tips and tricks you can use in your own dashboards (and you probably heard about something similar for Salesforce itself – DreamHouse App or Northern Trail Outfitters).

Some cool things

The conversational analytics (need to be enabled in Salesforce setup) just starts but can save time as you don’t really need to think to much how to do it but only how you want to formulate your question.

Use compare tables so you can quickly update and add columns, don’t be afraid to switch to SAQL, tweak the code and switch back. Also the SAQL editor in lenses is now-a-days the only place how to check whether you don’t have any type in your code. And don’t forget you can query SOQL directly from Salesforce, if you really need.

Binding is awesome for making the dashboard really dynamic – show me only clients who do more than 80 % of average client, all those bindings with static steps to change the dimensions of graphs or even their type or color. It is just beautiful and can save tons of different graphs.

Multivalue picklist work well here – in Salesforce it is evil, good to know it isn’t the case here. And you can also augment the records, so all the childs are flatten into the parent (kind of).

Security is a bit of tricky as it can replicate the Salesforce security just up to 2000 records for a user and works only on some objects – that was my biggest surprise. You have to use Security Predicates which aren’t that flexible and powerful and it is also lot of work.

EA Encryption requires Shield and when you have it your data are encrypted in rest and transit.

When will be some other Camp Fire?

All in all, these were greatly invested two days. Would be awesome to attend some other camp fire – I can think about CPQ or Field Service 🙂


Leave a Reply