1 Supporting Sakai: Lessons from the University of Michigan Sean DeMonner Jeff Ziegler Usability Support and Evaluation Lab
2 Usability, Support & Evaluation Lab, Digital Media Commons, University of Michigan What we’d like to share with you Brief history, scale and scope of the implementation Support group structure Descriptive statistics related to supporting a large scale Sakai implementation Support metrics and satisfaction indicators Common issues, and strategies for addressing them Best practices in support and lessons learned
3 Usability, Support & Evaluation Lab, Digital Media Commons, University of Michigan Overview - History Sakai in use at U-M for about three years Successfully migrated from legacy CMS to “CTools” CLE in Summer 2005 Currently running Sakai 2.1.2
4 Usability, Support & Evaluation Lab, Digital Media Commons, University of Michigan Overview - Scale Total existing sites 10,760 Course sites 7,217 Project sites Winter 06 Statistics 2,806 Course sites in use 44,740 unique users 12,700 daily users (avg) 4,190 concurrent users at peak
5 Usability, Support & Evaluation Lab, Digital Media Commons, University of Michigan Overview - Scope and phone support available during standard business hours Remote support Weekends Evenings (when warranted) 24x7 system availability except maintenance windows
6 Usability, Support & Evaluation Lab, Digital Media Commons, University of Michigan Usage Statistics - NetTracker
7 Usability, Support & Evaluation Lab, Digital Media Commons, University of Michigan Daily Usage Pattern
8 Usability, Support & Evaluation Lab, Digital Media Commons, University of Michigan Weekly Usage Pattern
9 Usability, Support & Evaluation Lab, Digital Media Commons, University of Michigan Support Structure Two-tiered Tech Support Model 1 st Tier – 90 hours/wk temporary staff Front line technical issues 2 nd Tier FTE professional staff Policy issues Issues escalated by 1 st Tier Training and Documentation 1.25 FTE professional staff.5 Staff Training.75 Help Documentation
10 Usability, Support & Evaluation Lab, Digital Media Commons, University of Michigan Support Structure Quality Assurance 30 hours/wk temporary staff Local Unit Support CTools Affiliate program Department IT staff Sites computing consultants
11 Usability, Support & Evaluation Lab, Digital Media Commons, University of Michigan User Training Ongoing, with a pre-term peak Instructor focused Group/department presentations preferred Low-stakes student introduction Coordination with release timing desirable
12 Usability, Support & Evaluation Lab, Digital Media Commons, University of Michigan Documentation Do not assume “no one reads the docs” 16% of faculty & 7% of students list docs as “most effective way to get help” Include docs in localization effort Occasional request for manual Online and task based PDF “chapters” Flash visual tutorials are being implemented
13 Usability, Support & Evaluation Lab, Digital Media Commons, University of Michigan Quality Control Internal staff tests local instance Risk based testing Test integration points with campus systems Test across campus browser pool 1-2 weeks preferred; semi-formal reports; go/no-go meetings Cross-test w/Sakai Operations Group runs load testing Evolution of protocol Automated load testing
14 Usability, Support & Evaluation Lab, Digital Media Commons, University of Michigan Tech Support Performance Metrics Number of support requests (91%) Phone (9%) Response Time Number of touches Satisfaction data from annual survey Daily and weekly status reports
15 Usability, Support & Evaluation Lab, Digital Media Commons, University of Michigan Issue Tracking FootPrints Electronic queuing mechanism Feeds metrics gathering and reporting Issue classification:
16 Usability, Support & Evaluation Lab, Digital Media Commons, University of Michigan Support Requests - LS&A mandate Fall site creation Winter site creation
17 Usability, Support & Evaluation Lab, Digital Media Commons, University of Michigan Support Requests - W06
18 Usability, Support & Evaluation Lab, Digital Media Commons, University of Michigan Response Time Service level goal: 2 hours or less Business hours constrain this goal *On average* response of 15 min. or less Peak times see increases in wait times Daily “clear the queue” push
19 Usability, Support & Evaluation Lab, Digital Media Commons, University of Michigan Touches Per Ticket N= 5518; does not include site creations These numbers should see some improvement due to “up front” requests for user info
20 Usability, Support & Evaluation Lab, Digital Media Commons, University of Michigan 2006 Support Survey - Instructors 1,357 respondents 33% have contacted support 18% report is “most effective way” they get help Instructors are “very satisfied” with support services (4.21 out of 5 quality rating)
21 Usability, Support & Evaluation Lab, Digital Media Commons, University of Michigan 2006 Support Survey - Students 2,485 respondents 15% have contacted support 36% report “keep trying on my own” is the most effective way to get help Students are “moderately satisfied” with Support Services (3.7 out of 5 quality rating)
22 Usability, Support & Evaluation Lab, Digital Media Commons, University of Michigan 2006 Support Survey - Comparison
23 Usability, Support & Evaluation Lab, Digital Media Commons, University of Michigan Support Request by Feature
24 Usability, Support & Evaluation Lab, Digital Media Commons, University of Michigan Status Reports Daily and weekly communications
25 Usability, Support & Evaluation Lab, Digital Media Commons, University of Michigan Voice of the Customer Customer feedback informs organization:
26 Usability, Support & Evaluation Lab, Digital Media Commons, University of Michigan Voice of the Customer Policy Issues often come through Support:
27 Usability, Support & Evaluation Lab, Digital Media Commons, University of Michigan Common Issues Sites created before registrar data is available Maintenance windows Large files & upload/download problems Site reuse vs. creating new
28 Usability, Support & Evaluation Lab, Digital Media Commons, University of Michigan Common Issues Integration with immature or unsupported services/browsers Lack of Auditing Tools Distributed product delivery (e.g. Identity management, File systems, Registrar data)
29 Usability, Support & Evaluation Lab, Digital Media Commons, University of Michigan Common Issues Managing large resource collections Managing peak support loads
30 Usability, Support & Evaluation Lab, Digital Media Commons, University of Michigan Lessons Learned Training is key Low-stakes introduction to tools Use a ticket tracking system Multi-tiered support staffing works well Hire and train Tier 1 staff early (July for Fall)
31 Usability, Support & Evaluation Lab, Digital Media Commons, University of Michigan Lessons Learned Distributed support model (Affiliates in units) Escalation of Issues / Communication Flow Advisory committee(s) / Policy making bodies Embrace the workaround Establish Support Accounts and use Jira Combine MOTD with Known Issues
32 Usability, Support & Evaluation Lab, Digital Media Commons, University of Michigan Lessons Learned Know browser targets and test local instance Migration never ends “Eat your own dogfood” Changing expectations of online systems There’s no substitute for talking with customers, attending trainings, etc. Support is the ear of the organization and should “have a spot at the table”
33 Usability, Support & Evaluation Lab, Digital Media Commons, University of Michigan Questions? Sean Jeff
34 Usability, Support & Evaluation Lab, Digital Media Commons, University of Michigan What else? FootPrints demo NetTracker demo Selecting support staff
35 Usability, Support & Evaluation Lab, Digital Media Commons, University of Michigan Abstract Sakai has been in use at the University of Michigan for the past 3 years, serving tens of thousands of customers. Hear the in and outs of supporting a large scale Sakai deployment from the folks who answer the phones and respond to the . Description: This presentation will review the experiences of the team supporting Sakai at the University of Michigan, including: - brief history, scope and scale of the implementation - support group structure & relations with other groups (Development, Operations, QC) - descriptive statistics related to supporting a large scale Sakai implementation - common issues and strategies for addressing them - best practices in support and lessons learned - support metrics and satisfaction indicators Attendees will come away with quantitative and qualitative information related to supporting a large scale Sakai implementation.