Operational Analytics
-
Upload
bi-leadership-forum -
Category
Technology
-
view
601 -
download
1
description
Transcript of Operational Analytics
© TechTarget
Operational Analytics Benchmark Report Results
December 2013
2© TechTarget
BI Framework 2020
Analytics Intelligence
Continuous IntelligenceCont
ent I
ntel
ligen
ce
Data Warehousing
Ad hoc query, Spreadsheets, OLAP, Visual Analysis, Analytic
Workbenches, Hadoop
Analytic Sandboxes
Event-driven
Reports and Dashboards
MAD Dashboards
Data Ware-housing
End-User Tools
Event-Driven Alerts and Dashboards
Ad hoc SQL
Dashboard Alerts
Event detection and correlation
CEP, Streams
Analytic Sandboxes
Design Framework
Architecture
Reporting &
Analysis
Excel, Access, OLAP, Data mining, visual exploration
Keyw
ord
sear
ch, B
I too
ls,
Xque
ry, H
ive,
Java
, etc
.
Map
Redu
ce, X
ML
sche
ma,
Ke
y-va
lue
pairs
, gra
ph
nota
tion,
etc
.H
DFS,
NoS
QL
data
bses
Business Intelligence
3© TechTarget
Two Worlds of Operational Analytics
Batch-loaded Data
WarehouseMini-batch fed
Data Warehouse
Trickle-fed DW with CDC
Complex Event Processing
Stream-based Processing
Update Cycle
DW Architecture
Non-DW Architecture
Days Hours Minutes Seconds Milliseconds Microseconds
4© TechTarget
Definition
● Operational analytics analyzes data on the fly. Real time data "streams" from multiple systems into an analytical engine without landing to disk or a data warehouse. The analytical engine monitors operational processes in real time, displaying activity and trends on an interactive dashboard. When data exceeds predefined thresholds or matches a rule, the engine can take automated actions, such as alerting users, executing a lookup, triggering a worfklow, executing a script, delivering a page, or updating a database.
© TechTarget 5
Respondent departments
Key Takeaways
• It’s interesting to note that 50% of buyers operate outside the IT department
• We continue to see growth in the number of users who have an IT title but who are more closely aligned with the business (21%)
With which part of the organization do you more closely align?
I'm in the IT department
51%
I have an IT role outside of
the IT de-partment
21%
I'm in a business
department not related
to IT28%
© TechTarget 6
Operational analytics adoption
Key Takeaways
• One-third of respondents have either “fully” or “partially” deployed operational analytics
What is the status of operational analytics at your organization?
No plans
Under consideration
Under development
Partially deployed
Fully deployed
26%
25%
18%
22%
10%
© TechTarget 7
Build or buy operational analytics
Key Takeaways
• Of those who have “fully” or “partially” deployed operational analytics, 57% have both built and bought their system
• Traditionally, companies build operational analytics system but there is a shift to buy full-fledged systems. This 17% will rise.
Did you build or buy your operational analytical platform?
Built
Bought
Both
27%
17%
57%
© TechTarget 8
Scope of operational analytics deployments
Key Takeaways
• Most operational analytics implementations are guided by the corporate IT department and integrate data from applications
Which best describes the scope of your operational analytic deployment(s)?
Enterprise
Business unit
Departmental
Inter-enterprise
60%
21%
13%
5%
© TechTarget 9
Operational analytics functional areas
Key Takeaways
• The top areas that use operational analytics are those that generate a lot of data on a daily basis and benefit from monitoring its predefined rules/functions
• These include Operations, Finance and Sales
What functional areas use operational analytics software?
Operations
Finance
Sales
Marketing
IT
Service
Supply chain
Risk
Product management
E-commerce
Logistics
Manufacturing
Other
65%
51%
50%
46%
42%
39%
24%
23%
21%
21%
21%
17%
6%
© TechTarget 10
Operational analytics primary users
Key Takeaways
• Operational analytic users are equally split between analysts and casual users
Who are the primary users of your operational analytics software?
Business analysts
Casual users
Application developers
Statisticians
45%
44%
6%
5%
© TechTarget 11
Operational analytics future plans
Key Takeaways
• 73% expect their company to expand deployment of operational analytics
• This is a strong endorsement that gaining insight using the freshest data possible delivers strong business benefit.
What are your future plans for the deployment of operational analytical tools?
Expand deployment
Maintain, but not expand
Decrease deployment
Other
73%
23%
1%
4%
© TechTarget 12
Operational analytics engine data feed
Key Takeaways
• The biggest data source is the data warehouse
• These results suggest most companies are delivering near real-time data, instead of real-time data using a CEP or ESP system
Which types of data feed your operational analytic engine?
Data warehouse data
Service or call center data
Point-of-sale or sales data
Local files (e.g., Excel, CSV)
Network data
Call detail records
Server logs
Email data
Trading or financial data
Clickstream data (e.g., Web logs)
Social media data (e.g., Twitter, Facebook)
Sensor data
Claims or warranty data
Hadoop/NoSQL data
Other
59%
45%
42%
39%
34%
34%
27%
27%
24%
24%
20%
19%
17%
15%
11%
© TechTarget 13
Operational analytics engine data sources
Key Takeaways
• Since the respondent pool runs operational analytics using a near real-time data warehouse, it’s not surprising that a majority of respondents cite more than six data sources.
How many sources of data does your operational analytic engine combine in your primary application?
0
1
2
3
4
5
6-10
0%
8%
9%
16%
10%
8%
50%
© TechTarget 14
Operational analytics engine data throughput rate
Key Takeaways
• More than two-thirds support throughput rates of more than 100 records per second, with 10% recording more than 100,000 records per second
What is the data throughput rate on average?
< 1 record per second
< 10 records per second
< 100 records per second
< 1,000 records per second
< 10,000 records per second
< 100,000 records per second
>100,000+ records per second
5%
13%
13%
27%
18%
14%
10%
© TechTarget 15
Operational analytics engine rule creation
Key Takeaways
• Not surprisingly, business analysts and business users create the rules for governing how the operational analytical engine manipulates data
Who creates the rules that govern how the operational analytical engine manipulates data?
Business analysts
Business users
Application developers
Data scientists
IT administrators
Statisticians
Other
53%
50%
37%
24%
24%
16%
5%
© TechTarget 16
Operational analytics engine data output
Key Takeaways
• Given the data warehousing platform, it’s not surprising that the most common output of an operational analytic system is a real-time dashboard (74%)
What is the output of the operational analytics engine?
Real-time dashboard
Alerts via Web, email or pager
Database updates
Workflow
Triggers (scripts)
New queries
Recommendations/offers
Trouble ticket
Other
74%
53%
46%
37%
27%
24%
22%
20%
4%
© TechTarget 17
Operational analytics supporting technologies
Key Takeaways
• More than three-quarters (76%) cited business intelligence tools as the most commonly used technology in an operational analytic system.
What technologies do you use in conjunction with operational analytics?
Business intelligence tools
Analytical databases
Data mining tools
Specialized "operational analytics" tools
In-database analytics
Rules engines
OLAP tools
Open source tools
Complex event processing engines
Streaming engines
Hadoop/HBase
Other
76%
47%
43%
42%
39%
33%
32%
27%
17%
11%
10%
9%
© TechTarget 18
Operational analytical applications governing rules
Key Takeaways
• Respondents apply a mix of Boolean and statistical rules in their operational analytics systems to automate alerts and actions
Which best describes the rules that govern your operational analytical applications?
Boolean
Statistical
Both above
20%
21%
68%
© TechTarget 19
Operational analytics software vendors
Key Takeaways
• More than one-third use Oracle, followed by IBM, SAS, open source software and Informatica
• These results show there is a clear need for Vitria to enhance their market presence among these buyers.
Which vendors supply you with operational analytics software?
Oracle
IBM
SAS
Open Source (Flume, Storm, Kafka)
Informatica
SQLStreams
Sybase (SAP)
Splunk
HP
Tibco
Streamworks
InfoChimps
ZoomData
Splice
Vitria
Tervelo
Other
35%
27%
22%
21%
17%
11%
11%
11%
7%
7%
6%
2%
2%
1%
1%
0%
42%
© TechTarget 20
Operational analytics engine benefits
Key Takeaways
• Respondents cited a litany of benefits, including improving operational efficiency and working more proactively
• Given the high scores across the board, it’s clear that operational analytics creates a positive ripple effect throughout an organization
To what degree does your operational analytical engine deliver the following benefits?
Improve operational efficiency
Work more proactively
Detect problems quickly
Increase competitiveness
Improve data quality
Increase business transparency
Improve customer experience
Reduce costs
Automate actions
Increase revenues
50%
44%
43%
41%
40%
35%
34%
31%
30%
25%
© TechTarget 21
Operational analytics engine challenges
Key Takeaways
• Respondents cited “sourcing data” and “defining rules for analysis and actions” as the top challenges.
• Surprisingly, “scalability” and “performance” were the least cited challenges
What challenges have you faced implementing operational analytics?
Sourcing - Capturing data from multiple, complex systems
Complexity - Defining rules for analysis and actions
Scalability - Ingesting high volumes of data
Performance - Maintaining performance as query and data complexity increase
Funding - Getting executives to fund the installation or expansion of the software
Integration - Integrating tools with other information environments
Data Quality - Identifying and fixing data quality errors
42%
42%
25%
26%
36%
34%
32%
© TechTarget 22
Operational analytics software obstacles
Key Takeaways
• Respondents who have not implemented operational analytics cite that they “Don’t know enough about [it]”
• Since operational analytics is a newer discipline, it’s not surprising that a large percentage of respondents haven’t heard about it yet.
What prevents you from deploying operational analytics software?
Don't know enough about them
Our budget is tapped out
No need
We built our own
Other
Performance and scalability issues
Not enough value for the price
34%
28%
17%
14%
13%
12%
10%
23© TechTarget
Summary
● Operational analytics is an early adopter market.● Lots of headroom among the BI audience● BI audience using traditional BI technologies to satisfy
operational analytical applications and near-real-time information delivery.