Salesforce Einstein Analytics Platform - Review 2022

Salesforce Analytics, which is a relative newcomer to the cocky-service business organisation intelligence (BI) infinite, is a suite of several different products. First and foremost is the artificial intelligence (AI)-based Salesforce Einstein Analytics Platform (which begins at $75 per user per month). While its analytics are primarily aimed at mining customer and sales data, information from other sources tin as well be connected to the tool. Very few BI products these days connect to Salesforce. Editors' Choice IBM Watson Analytics is fully integrated in terms of capabilities. The just identify the latter will fall down for some is in its data visualization capabilities (more than on that subsequently).

Among other advantages, the doubling down on predictive analytics from this combo Watson-Einstein dance makes Einstein await like a bright performer. Just, alas, each dance partner must be judged on their ain repertoire. And that's what I'll exist doing here.

First, a look at the Salesforce Analytics lineup. Bones analytics are costless. For more advanced analytics, at that place'southward the Sales Analytics awarding for Sales Cloud, priced at $75 per user per calendar month. Then there's the Service Analytics app for Service Cloud, which is too priced at $75 per user per calendar month.

At the height of the lineup is the Salesforce Einstein Analytics Platform (as well priced, as stated before, at $75 per user per month), which includes sales and service apps plus a slew of other features. These features include the ability to analyze data from whatever source (upward to 100 million data rows is advertised but we shall see), the ability to create an unlimited number of custom apps, templates to build apps and dashboards, and unlimited online grooming.

I have covered Salesforce since its debut on the marketplace as a Software-as-a-Service (SaaS) pure play in an age when client relationship direction (CRM) was little more than than a glorified contact management production in a box. Now, just as then, Salesforce sets its sights on leading the market by doing things differently. But now is not and then, and this time Salesforce is not the first on the dance floor. Information technology is, however, but fashionably late as its Einstein is following simply a few months later IBM'southward Watson. Einstein is impressive and powerful and, when used with Salesforce information in item, it meets or exceeds the needs of roughly half of all enterprise users—the half who are focused on customers, marketing, and sales. It can be useful to other users as well; it's just that sales and marketing has always been Salesforce's sweet spot and it still is.

The Salesforce Einstein Analytics Platform uses motorcar learning (ML) merely non deep learning. It isn't truly cerebral computing (that is, AI) just as Watson isn't. Just Salesforce as a company continues to prove business acumen in a league of its own. The company is willing to forge partnerships (master among them the full Watson integration) to secure a straight path to enabling the rise of citizen data analysts—the holy grail of the cocky-service analytics manufacture. Still, the Salesforce Einstein Analytics Platform has its limits and points of confusion.

Salesforce Einstein Analytics - Landing page

Getting Started

Y'all find your way to a trial business relationship by selecting and following one trail from many. A "trail" in this case is Salesforce speak for a "learning path" or a "learning trail. That's why they call their partners "Trailblazers." I constitute this path to exist initially cumbersome and a time drain. I adopt to click a "Trial Account" button or registration page and become on with the piece of work. It didn't assist that I even asked the spokesperson to send me credentials and a link for a trial account for this review, and she sent me 2 or three trails to follow instead.

The spokesperson finally told me that following a trail would lead to the formation of a trial account. While she was very nice and obviously trying hard to be helpful, this felt similar a far more than disruptive setup than is warranted—especially if you're aiming for business organization users (aka citizen information analysts) as Salesforce claims it is.

Nonetheless, I picked a trail. Rather, I picked the Salesforce trailhead named "Einstein Discovery Basics" that the helpful spokesperson aimed me at, saying this trailhead is "meant to get business users up and running quickly and on their own, without the help of a data annotator." The easiest trailhead, presumably.

Lo and behold, the trailhead was helpful indeed. I easily and rapidly set up the account and imported data by following the piece of cake-to-understand instructions on the trailhead. Then, okay, the trailhead concept works. But I withal don't understand how users are to find and and so determine which trail to follow to go where they want to exist. I searched online for "Salesforce trailheads" and got a long list. Nevertheless, I accept no thought as to which to pick if in that location isn't a handy spokesperson nearby to tell you.

One word of circumspection, to follow me down this trailhead: Delight note that yous cannot utilize an existing Developer Edition (DE) org that you have in Salesforce. Instead, you lot must sign up for a new ane (the instructions on this trailhead will tell you how) because you lot need the limited analytics license in the new DE org.

As for my feel, several bumbles and fumbles afterwards, I managed to navigate to the landing page in the new DE org. At that place, I clicked on the app launcher, and and then searched for and clicked to launch the Einstein Discovery app. Merely nothing happened until I immune popular-ups on my browser. So be sure to practise that.

Salesforce Einstein Analytics - Data exceeds maximum warning

Once the Einstein Discovery app was open up, I clicked the tile labeled "CSV" and loaded my data (the same data sets I used to review competing products in this category) easily plenty. But and so the Salesforce Einstein Analytics Platform responded with "Exceeds maximum rows allowed (91,980 constitute). Only the terminal 50,001 rows will be candy." The deplorable matter is, it gave me that alert later on I had loaded only two files of approximately 30 rows. The other sad affair is, Salesforce advertises Einstein as capable of treatment "upwardly to 100M data rows," simply this is conspicuously not what this screenshot shows (come across above). Einstein also warned me of "Big numeric values found in the column." To which I say, "Aye, and?!" Grrr, Einstein wanted that "fixed" to its limits, too.

It noted other problems with the data which amounted to my taking standard information prep deportment. One such problem was that date and time were both present in the same column. Either of those values could easily be ignored or the data cleaned upward. No biggie at that place. The limitation on information size is a far bigger concern.

This presents a huge ruddy flag to me. But it may exist less or more alarming to y'all, depending on the size and/or configuration of the information on which you're trying to work. If you lot're just using Salesforce data, then it'due south all skillful considering the two are meant to work together! If you're using other data, also, then all-time look at that upshot closely first.

For the sake of this review, I chose to use the sample data set provided by Salesforce in the trailhead rather than reconfigure my data sets to reduce the number of rows to fit this arrangement'due south limits (there'southward just so much time-suck one should accept to suffer in the process of picking or rating a production, don't you recall?) On with the review, so.

Salesforce Einstein Analytics - Results dashboard

The Discovery Process

Similar to several BI apps, especially SAP Analytics Cloud, the Salesforce Einstein Analytics Platform prompts you to create a story. That'due south a good approach considering humans ingest and retain stories meliorate than numbers equally a full general rule. Tiles are presented to assistance me configure everything the way I want it before I hit the "Create Story" push. (Yep, this is the build-a-query stage.)

One time I do hit that button, the system takes simply a few seconds to clarify the data and return a storied dashboard to me. On that dashboard is a lot of information (more than than I can show you lot in a single screenshot) and some text specifying what happened, according to this data set.

Further, on the pull down menu nether "Recommended Insights on" which reads hither every bit "What happened" are choices to notice "What changed over time," "Why it happened," "What could happen," and "How tin I meliorate it?"

Those are, of course, exactly the questions a seasoned data analyst would ask. Just here it'southward as well packaged in a way that a newbie, citizen information annotator can hands make the query. This entire discovery procedure aimed at being useful and usable to anyone, at well-nigh any skill level, is but brilliant.

This where the Salesforce Einstein Analytics Platform edges close, just not at, the head of the grade. Watson all the same owns this space because of its power to analyze larger data sets of both structured and unstructured information and to query in tongue beyond the general (just very of import) questions Salesforce offers here. This is a far cry from the more difficult-to-learn, SQL-oriented BI platforms, such as Chartio.

However, if you lot're working with structured data in a format friendly to the Salesforce Einstein Analytics Platform (sentry out on the number of rows, for example) and are primarily seeking insights related to sales, marketing, and client relationships/customer experience, then you're probably going to love the Salesforce Einstein Analytics Platform.

Salesforce Einstein Analytics - Create a story

Data Visualizations

I institute the "Play Summary" choice on the dashboard to be a unique characteristic, but I'chiliad not sure how many users would want the Salesforce Einstein Analytics Platform to read the results to them. You can besides "personalize" the story, which is a thumbs upwardly or thumbs downwards vote on line items in the study. It's non articulate equally to what stop, though. Nevertheless, you can share and export the story, also, directly from the dashboard and that's always helpful.

As to visualizations proper, well, that's a different story than yous see in other analytics apps. In short, y'all don't really choose visualizations in the Salesforce Einstein Analytics Platform. Y'all modify the query and Einstein automatically picks the visualization that best displays the answer.

Earlier you label this a creative injustice, call back that the end goal is to provide analytics that are usable by and useful to the business masses—well-nigh of whom it can exist causeless care little about visualization choices. Simply if yous really care about having a full palette of visualizations at your fingertips, then you're not probable to be happy with the Salesforce Einstein Analytics Platform'due south automated, basic displays and reports.

While that might limit Einstein'due south potential for some, imagine taking all of that and integrating it fully with IBM Watson. If you've got the budget, and so Salesforce and IBM take done a deal that enables this choice. The result is a combined product that's damn tough to shell by anyone else in the field. Lookout out world: These two partners are low-cal-years ahead in the data democratization race.