Microsoft Power Stack 2019 [Power BI, Excel, Azure & Friends]

Making Sense of
Microsoft’s Sprawling, Shifting
Business Intelligence Offering
Power BI & Beyond
September 2017 Overview

What’s in this presentation: bringing clarity to
crowded field of products, architectural choices
• Microsoft’s BI offering
• Historical perspective and dynamics
• Strengths & Weaknesses
• Shared Assets: Power Query, Power Pivot
• Products & Licensing
• SKUs & Pricing
• Components & Integration Points
• Hope you liked Inception
• Development Options & Languages
• DAX, M, MDX, R, T-SQL…
• Clarifying seemingly competing options
• Client BI tools: Power BI vs. Excel
• Cloud or On premises: Power BI Desktop vs. Cloud services vs. Power BI Report Server vs. PBI Gateways
• Reporting UIs: Power BI Reports vs. Dashboards
• Connection types: Import vs. DirectQuery vs. Live Connection
• Data models: Tabular vs. Multidimensional
• BI approach: self-service and/or Enterprise IT
• Plenty of supporting hyperlinks (in slides and footer notes)

How We Got There
And why should I care?
Microsoft has been in the BI space for a long time with a “traditional IT”
focus
• Think SQL, OLAP, multi-dimensional, enterprise IT, data warehouses
Circa 2007-14 MSFT seemed to leave the world move to SaaS without
putting much of a fight
• They let the likes of Salesforce, Google, or Tableau, make them look old and irrelevant
And lots of moving parts took time to mature and gel
• Power Pivot was first launched in 2009 but was relegated to add-in status until Excel 2016
• Power View/Silverlight dead-end (mobile anyone?)
But since 2015 there has been a strong move to inexpensive yet
powerful desktop + cloud + mobile services for business users
• Think tabular, spreadsheets, SaaS, end-user friendly, within reach of smaller teams without
requiring IT
Come 2017: still work in progress, but a compelling offer
• Strong momentum behind Power BI (PBI) makes it unlikely they will drop the ball, so worth
investing into its vision
• Good old Excel has become a formidable yet more accessible tool, aka “Modern Excel”
“Windows Only”

Exec Summary: Lots to Like, Some Drawbacks
Strengths
• The broadest BI/DB stack of any
vendor, with individual components
strong in their own right.
• Builds up on established leading
products (Excel, O365, SQL Server,
Azure).
• Vibrant product development, iterative
and interactive.
• Strong user community, support,
learning material.
• Easy and cheap to get started.
Weaknesses
• Hard to keep up with all the moving
parts and updates!
• Still somewhat Microsoft-centric, but
much less so than in the past.
• Not as good as Tableau for more
sophisticated interactive charts.
• A lot of stuff still in beta/incomplete.
Needs another 12-18 months of
maturation.
• Licensing can get complicated, some
market segments feel neglected.

Components of the “Power Stack”
Self-service BI, Workflows, Apps
Software Platform What General availability Notes
Excel 2016 Windows Spreadsheets
Data models
Viz
September 2015 Until Excel 2013 (included), Power Query
and Power Pivot were separate add-ins
Power BI
Desktop
Windows Data models
Viz
July 2015 Authoring tool formerly known as Power
BI Designer
Power BI Service Cloud Viz July 2015 SaaS
Power BI Mobile iOS, Android,
Windows
Viz July 2015 Native mobile apps
PBI API
PBI Embedded
Cloud Integration July 2015
January 2017
Embedded merged into Premium in mid-
2017
Flow Cloud, iOS,
Android
Inter-app workflows
& integration
October 2016 Think SaaS glued with Zapier, IFTTT, i.e.
“Lotus Notes for the cloud.”
PowerApps Cloud
Windows
Visual app designer October 2016 Think Force.com “low code”, i.e.
Access/VB for the cloud. PowerApps
plans include Flow.
OneDrive for
Business
Cloud
Windows
Mobile
Cloud file sync February 2014 Think Sharepoint SaaS. Allows auto-
refresh of desktop Excel files into cloud
PBI.

Sub-components of the “Power Stack”
Two common core technologies and user experiences
Subcomponent /
Software
Function Excel Power BI SQL Server Azure Common
Data Service
Power Query ETL Y Y SSDT 2017 Yes
Power Pivot Data modeling Y Y SSAS Analysis Services

Licensing
This is Microsoft, ergo cannot be simple
Software Platform Office 365 Dynamics 365 Separate license Notes
Excel 2016 Windows (no
PQ/PP in Mac
version)
• (Limited PQ/PP in Business plans)
• ProPlus ($12/u/mo)
• Enterprise E3-E5 ($20-$35/u/mo)
n/a Office 2016 perpetual license (i.e. not a subscription) • O365 business plans
• New ways to get the Excel
business analytics features
you need
Power BI Desktop Windows n/a n/a Free download
Power BI Service Cloud Enterprise E5 n/a • Free
• Pro ($9.99/u/mo)
• Premium P1/P2/P3 ($5K+/mo)
• Pricing
• Premium cost planner
Power BI Report
Server
Windows Server
On premises
n/a n/a • Premium P1/P2/P3 ($5K+/mo)
• SQL Server Enterprise Edition with Software Assurance
Introducing PBI RS
Power BI Mobile iOS, Android,
Windows
n/a n/a Free downloads
PBI Embedded Cloud n/a n/a Premium EM1/EM2/EM3 ($625+/mo) Embedded pricing
Flow Cloud, iOS,
Android
• Business Essentials/Premium
• Education / Education Plus
• Enterprise E1/E2, E3/E4, E5, K1
(restricted)
• Enterprise: Sales, CS, Ops, FS, PSA,
TM, Plan 1/2
• Business: Financials, TM
• CRM
• Free: 750 runs, 15mn checks
• (O365 base plan: 2K runs, 5mn checks)
• Plan 1: 4.5K runs, 3mn checks ($5/u/mo)
• Plan 2: 15K runs, 1mn checks, admin ($15/u/mo)
• All PowerApps plans
Pricing
PowerApps Cloud, Windows • Free Community plan
• (O365 base plan)
• Plan 1: premium connectors, use CDS apps ($7/u/mo)
• Plan 2: admin + build CDS apps ($40/u/mo)
• Pricing
• Licensing Overview
OneDrive for
Business
Cloud
Windows
Mobile
• Business / Essentials ($5/u/mo) /
Premium: Plan 1
• ProPlus: Plan 1
• Enterprise E1: Plan 1
• Enterprise E3-E5: Plan 2
• Plan 1: 1TB/user ($5/u/mo)
• Plan 2: 5+TB/user, data loss prevention ($10/u/mo)
Pricing

ETL
Power Query
Model
Power Pivot
Data
JSON
CSV, Excel
Streaming datasets
Sources
Databases
APIs/SaaS
Files
View
Pivot Tables
Pivot Charts
(Power View)
View
Reports
Dashboards
Other apps/sites
Get, Combine, Transform, Model, Visualize
& Share Data with the Power Stack
public
web

Connection Inception: How Stack Pieces Interact
Source(s) Destination What How Notes
Excel Desktop PBI Service Queries
Workbooks
Data Catalog
Publish to PBI
PBI Pro subscription needed.
PBI publisher for Excel needed.
Excel Desktop PBI Desktop Queries, Data,
Data models,
worksheets
Import Power Query queries, Power Pivot models and Power View worksheets.
Excel Desktop
PBI Desktop
PBI Service Data model Local/LAN via PBI
Gateway
Need a gateway to refresh datasets that get data from on-premise data sources. PBI Pro
subscription needed.
Excel Desktop
PBI Desktop
PBI Service Data model Cloud via OneDrive,
Sharepoint Online
Any data you’ve loaded into your file’s model is imported into the dataset and any reports you’ve
created in the file are loaded into Reports in PBI. Changes to your file on OneDrive will be updated
in PBI too, usually within about an hour.
PBI Desktop Excel Desktop Data model Analyze in Excel, Use
PBI Desktop as a
local server
Since users will need to refresh the dataset, and refresh for external connections is not supported
in Excel Online, it’s recommended that users open the workbook in the desktop version of Excel on
their computer.
Excel Desktop, PBI
Desktop, PBI Service
3rd-party Apps Dataset PBI REST API You can import pbix and xlsx files stored on OneDrive via the API.
PBI Service PBI Desktop, Excel
Desktop
Dataset Connect to datasets
(live connection) ,
Analyze in Excel
Reusing datasets
PBI Service 3rd-party Apps Viz PBI Embedded
PBI Service 3rd-party Apps Data alerts Flow Notification triggers.
PBI API PBI Service Datasets, reports,
dashboards
API calls
PowerApps PBI Apps PA/PBI interop Embed PowerApps within PBI dashboards + Embed PBI tiles into PowerApps apps
Common Data Service PBI, PowerApps, Data model Common Data Excel support via add-in.

Confusion Infusion: MSFT Loves to Move, Rename Features
Excel 2010: It’s Coming from the SQL Team, So Let’s Hide It
• PowerPivot introduced as an Excel add-in, common engine with SSAS Tabular
• In-memory engine introduced as Vertipaq then rebranded to xVelocity, because why not?
Excel 2013: Moving Sideways
• Power Pivot (now with a space, no product name can be left untouched!) remain as an add-in
• Introduction of Power View and Power Map based on Silverlight dead end
• Introduction of Data Explorer then rebranded as Power Query (notice a trend here?)
Excel 2016: Finally Embracing the Power
• Power Pivot is its own ribbon tab but also accessible from Data > Manage Data Model
• Power Query renamed as Get & Transform
• Power View hidden though still available, Power Map renamed to 3D Maps
• Old way to get external data continued to coexist with Get & Transform until March 2017
Power BI Desktop: There Can Be Only One Power Something
• Power Query found as Home > Get Data (but some documentation does talk about “Power Query”)
• Power Pivot found as the Modeling ribbon tab (let’s not name it “Power Pivot” because it might look familiar to Excel users)
PBI Service: Office Or Not?
• “Power BI for Office 365” was an older version of the service using Excel (with Power View/Map) for authoring and was deprecated at the end of 2015. But
powerbi.com user management continues to rely on O365.
Azure: What’s Up with That?
• Retirement of DataMarket, Data Services, Reporting Services
• Planned merger of PBI’s data catalog with Azure Data Catalog

Great Momentum, But Uneven Features Between Products
• Streaming datasets launched Jan. 2017 in PBI Service, not available in PBI Desktop
• Can’t print or export to PPT from PBI Desktop
• Some mapping features only available in Desktop (ARCGIS) or Service (aerial basemap)
• Many online services and CDM as a native Source in PBI but not (yet?) in Excel 2016
• Calculated Tables, Bidirectional cross filtering available in PBI Desktop, not in Excel 2016
• Some UX pieces offer better QoL to author DAX in Excel than Power BI
• OneDrive listed under Get Data > Files in PBI Service, but not Desktop
• Some objects such as Hierarchies recognized in Excel’s data models but not in PBI Desktop
• Excel imports/exports M queries to ODC files but PBI Desktop doesn’t
• No Dashboards and no natural language Q&A in PBI Desktop / PBI Report Server
• Need to install a separate “PBI Desktop for Report Server” for On Premises work
• SSDT in VS2017 not available at launch
Don’t despair: lots of monthly updates, many parts still in beta/preview (e.g. most SaaS connectors)

Even More Caveats: “Pulling a Microsoft”
Some questionable choices
More Open but Still Somewhat MSFT-centric
• Authoring limited to Windows and to a smaller extent browser
• No Mac desktop
• Flow/PowerApps connectors still focused on MSFT world
• SQL, Sharepoint, Azure
Abrupt & Enterprise-Focused Licensing Changes
• Changes to Free/Pro announced out of the blue within less than a month of their
enactment
• 1-year free Pro trial thrown in to help you stomach it
• PBI Premium starts at $5K/mo
• Makes sense only from 500+ users
• Forced migration from PBI Embedded (Premium Embedded pricing starts at $625/mo)
• Why tie PBI Report Server to PBI Premium and SQL Server Enterprise with Software
Assurance?
• SMEs and ISVs Might Want On-Premises PBI…

Clarifying the
Power BI offering

Microsoft Power Stack 2019 [Power BI, Excel, Azure & Friends]

Power BI Desktop vs. powerbi.com cloud service
Two concepts that look similar but complement each other
Functionality Desktop Cloud
Positioning Authoring tool for analysts Consuming & Sharing tool for end users
Data sets Yes Yes
Reports Yes Yes
ETL (Power Query) Yes No
Data modeling (Power Query) Yes No
Visualization Yes Yes
Data access based on row level security No, as a report author you get access to the whole data source Yes
Dashboards No Yes
Export to PPT No Yes
Print No Yes
Controlled sharing Not really
(you can share the whole .pbix file)
Yes

Reports vs. Dashboards
Two concepts that look similar but complement each other
Functionality Reports Dashboards
Multiple pages/tabs Yes No
Drill down/up, Expand, See Data Yes No
Publish to web Yes No
Slicers Yes No (unless you pin the whole page from a report)
Export to PPT Yes No
Print Yes Yes
Subscribe Yes Yes
Automatic refresh No Yes
Q&A No Yes
Alerts No Yes
Featured No Yes
Favorite No Yes
Sharing No Yes
Web/phone view No Yes

Power BI Report Server
Power BI on Premises, with Strings Attached
Functionality Powerbi.com Power BI Report Server
Licensing Flexible and
starts cheap
Inflexible and expensive
(enterprise only)
Updates Monthly 3x a year
(server and authoring client)
Authoring client Power BI Desktop Separate PBI Desktop version
Variety of data sources Yes No, only SSAS live connection
(more currently in August Preview)
Dashboards Yes No
(only reports)
R Visuals Yes No
On premises No Yes
(well, it’s got that going for it)

Mix & Match On Premises & Cloud
PBI Enterprise Gateway

Gateway: 1 Installer, 2 Modes
GATEWAY (PERSONAL MODE)
Mode (Choose on Install) On-premises data
gateway
Personal
installation
Cloud services it works with Power BI, PowerApps,
Azure Logic Apps,
Microsoft Flow
Power BI
Runs as an app for users who aren't administrators on the computer •
Runs as a single user with your credentials •
Import data and set up scheduled refresh • •
Support for DirectQuery to SQL Server, Oracle, Teradata •
Support for a live connection to Analysis Services •
Serves multiple users with access control per data source •
Monitoring and auditing for gateway and data sources Coming soon

Enterprise Offering Continues to Evolve in Parallel
Increasingly Integrates, Blends with Self-Service
• SQL Server Integration Services (SSIS)
• Enterprise ETL
• SQL Server Analysis Services (SSAS)
• Think of SSAS Tabular as “PowerPivot Enterprise Server”, i.e. modeling tool
• “Live Connection” option in PBI: no need to import data but faster than Direct Query
• PBI can query SSAS Tabular and SSAS MD
• Models can be imported from PBI Desktop
• SQL Server Data Tools (SSDT)
• Now supports DAX queries against SSAS
• SQL Server Reporting Services (SSRS) & PBI Report Server
• Focus on On-Premises, pixel-perfect/print static reports
• Paginated reports and Mobile reports can be pinned in PBI dashboards
• PBI reports can be put in SSRS dashboards
• PBI RS is basically a superset of SSRS
• SQL Server Management Studio (SSMS)
• Now has a DAX query editor

Tabular or Multidimensional: PBI + SSAS
Tabular Multidimensional
Available since SQL Server 2012 SQL Server 2000
Licensing SQL Express and above, PBI, Excel SQL Server Enterprise
Modeling constructs Relational: model, tables, columns OLAP: cubes, dimensions, measures
Languages DAX calculations, DAX queries, MDX queries MDX calculations, DAX queries, MDX queries
M Expressions Yes No
Calculated measures, columns, tables Yes Calculated members
Actions No Yes
Aggregations No Yes
Custom Assemblies No Yes
Custom Rollups No Yes
Default Member No Yes
Display folders No Yes
Hierarchies Yes Yes
KPIs Yes Yes
Many-to-many relationships No Yes
Writeback No Yes
Compression vs. original data size Can go down to 10% About 30%
Sources Largest variety Only relational databases
Database size MD used to be needed for the largest projects but Tabular is catching up

PBI-SQL Connection Types
Choose were storage and modeling happen
Connection type Data storage Data
modeling
Refreshing Notes
Import Data loaded in PBI’s in-
memory engine
PBI
• Scheduled refreshes
(every 15 minutes or
more)
• Limited to 1GB dataset and
10GB of data in cloud service
• Can’t change connection type
Direct Query
Data remains in the source
SQL database
• Data is pushed via the
Enterprise Gateway
• You can schedule cache
refreshes
• Pinned tiles are
automatically refreshed
every 10 minutes
• Can change connection type
• No Quick Measures
• No calculated columns
• Can’t create hierarchies
• Can’t connect to several
sources
• Can’t create relationships
between tables
• Can’t change connection type
• Most scalable solution
Live Connection
SSAS

Mix & Match On Premises & Cloud
Services galore
Service On premises Cloud
Database (OLTP), Data Sync SQL Server Azure SQL, Azure Data Sync
Data warehouse SQL Server Azure SQL Data Warehouse
Data Modeling SSAS Azure Analysis Services, PBI
ETL SSIS Azure Data Factory, PBI, Future Power Query SaaS stand-alone
Reporting PBI RS, SSRS PBI
And more… IaaS: SQL in Azure VMs
PaaS: Azure Storage, HDInsight, Data Lake, Event Hubs, Stream Analytics

BI & DB Languages
Sorting out the alphabet soup
Language Sinc
e
About MSFT Support Strengths / purpose
T-SQL (Transact
Structured Query
Language)
1974
(SQL)
Query language for relational databases. SQL Server Work with individual rows and row sets.
MDX (Multidimensional
eXpressions)
1997 Query language for OLAP databases. SQL Server, PBI Desktop Query data tubes. Powerful.
DMX (Data Mining
Extensions)
1999 Query language for SSAS data mining
models.
SQL Server, Excel Syntax looks like SQL.
DAX (Data Analysis
eXpressions)
2009 Data modeling language introduced with
PowerPivot.
SSAS, Excel, PBI, DAX Studio, DAX Editor Calculate dynamic aggregates with Excel-like function syntax.
Flexible and versatile.
M 2010 Functional language introduced with
Power Query.
Excel, PBI ETL
LINQ (Language
Integrated Query)
2007 Query language for .NET languages. Visual Studio: C#, F#, VB.NET
R 1993 Statistical and graphical language. Open
source but embraced by Microsoft.
Excel, PBI, SQL Server, Visual Studio, R
Server & Client
Stats (e.g. regression analysis), LOTS of extensions, good
charting.
Python 1991 Programming language popular with data
scientists
SQL Server 2017 The most general purpose language in this table. Good to
handle text.

Power BI Development Options
• Power BI REST API
1. Push datasets/tables/rows into PBI model in language of your choice
2. Trigger data refreshes
3. Embed dashboards in your app (now using the same API)
• Custom visuals
• TypeScript / CSS / D3
• Data Connector SDK / Cloud content packs
• PowerApps + Flow
• Form-based “no code” apps embedded in Power BI dashboards
• Visual workflows using PBI alerts as triggers; pushing data to PBI datasets
• R: ETL and Charting
1. Load an R script as a data source
2. Run an R script as a Power Query step
3. Plot an R visual (including using libraries such as ggplot2)
4. Create an R-based custom visual (including using Plotly)
• VBA
• Publish Excel workbook to PBI

The Complete Desktop Toolset for the MSFT BI Analyst
Most of that stuff is free!
“Self service”
• PBI Desktop
• Custom charts
• Content packs
• Themes
• Personal Gateway
• Excel 2016 (ProPlus, Office Insider)
• PBI Publisher for Excel
• Analytics Edge
• SQL Server Express
• SSMS (SQL Server Management Studio)
• Advanced Services (subset of SSRS, not sure
about SSIS)
• OneDrive
• DAX Studio
• R Studio
“Enterprise”
• PBI Desktop
• On-premises Enterprise Gateway
• SQL Server Dev Edition
• SSAS
• SSIS
• SSRS
• MDS (Master Data Services) for SSMS
• VisualStudio
• SSDT
• R Tools
• DAX Editor
• Power Update

What Will This Look Like in 2019?
Disrupting Tableau’s momentum?

About these slides
Tech + data to meet business goals
I’m an entrepreneur and consultant working at the intersection of
business and technology, with a background in sales and marketing.
I use Excel and PowerBI to integrate analytics from many sources such
as Google Analytics or Salesforce.com. Feel free to contact me if you
need help choosing your BI tools, planning your business intelligence
project, or creating your dashboards.
I’m maintaining and improving this presentation as I learn more and
the offering evolves. Did this presentation help you? How can I improve
it? Your feedback is appreciated.
To contact me:
• Blog
• LinkedIn
• Twitter
• olivier@needlestacker.com
Olivier Travers

Microsoft Power Stack 2019 [Power BI, Excel, Azure & Friends]

More Related Content

What's hot (20)

Similar to Microsoft Power Stack 2019 [Power BI, Excel, Azure & Friends] (20)

Recently uploaded (20)

Microsoft Power Stack 2019 [Power BI, Excel, Azure & Friends]

Editor's Notes