Dairy of Support Engineering Manager
1 like2,352 views
The document is a diary entry by Toru Takahashi, Support Engineering Manager at Treasure Data, detailing the evolution and performance metrics of their support engineering team from 2015 to 2018. It highlights improvements in customer support processes, the development of a support data platform, and challenges faced with communication and data management. Additionally, it outlines future plans for enhancing customer experience and support tools, while also announcing ongoing hiring efforts.
Dairy of Support Engineering Manager
- 1. Diary of Support Engineering Manager Toru Takahashi Treasure Data, Inc. サポートエンジニアNight vol.2 2018-02-07
- 2. Who AM I ? Profile - 髙橋 達 (Takahashi Toru) - Treasure Data, Inc. - Support Engineering Manager - Twitter: @nora96o / Github: toru-takahashi - Blog: http://blog.torut.tokyo/ Misc. - Tech. Support Engineer from the almost beginning of my career - 1st Tech. Support Engineer in TD - Support Engineering Manager from 2015
- 3. Treasure Data Support - Email - Online Chat - Support Form
- 4. Contents - Past / Present / Future - Defined Measurable KPIs for Support Performance - Built Support Data Platform - Improved a process for Customer Slack Channel - Multiple Components = Huge Logs
- 5. Past / Present / Future 2015 / 2017 / 2018
- 6. 2.5 years ago... What is Support Engineer in TreasureData
- 7. Treasure Data Support From 2015 to 2017/2018 Head Count from 2 to 5 to 8
- 8. Treasure Data Support From 2015 to 2017/2018 Head Count from 2 to 5 to 8
- 9. Treasure Data Support From 2015 to 2017/2018 Head Count from 2 to 5 to 8 (plan…)
- 11. Treasure Data Support From 2015 to 2017/2018 Avg Ticket/Month from 308 to 477 to 500~600?
- 12. Treasure Data Support From 2015 to 2017/2018 TreasureData Components 2015
- 13. Treasure Data Support From 2015 to 2017/2018 TreasureData Components 2017 (Treasure Reporting)
- 14. Treasure Data Support From 2015 to 2017/2018 TreasureData Components 2017 (Treasure Reporting)
- 15. Treasure Data Support From 2015 to 2017/2018 TreasureData Components 2017 (Treasure Reporting)
- 16. Treasure Data Support From 2015 to 2017/2018 TreasureData Components 2018 = Treasure CDP (Treasure Reporting)
- 17. Defined Measurable KPIs For Support Performance
- 18. How Do You Measure Support Performance? - Customer Satisfaction Score? - Number of Resolved Tickets? - Average Resolution Time? Ref. Top 10 Key Performance Indicators for Customer Service - But, we’re growing - Number of customers increases - Number of new tickets increases - New features are released every week - Difficult to find out “Stable/Controllable” KPI
- 19. Support Performance = Provide “Great Support” - Great Support → Make a customer our fan - Trying to measure factors of “Great Support” by the following KPIs - Satisfy 100% SLA (1st Response Time) - A ticket resolution rate within 7 days to 70% - Quick Resolve helps customers - Engineer Escalation Rate to under 15% - Related to quick resolve - Engineer focuses on development new features for customers - Avg. Response Time to within 1 business days (Stretch) - Updating a ticket frequently give customers a kindness
- 20. Built a KPI dashboard for Support Team
- 21. Built Support Data Platform
- 22. Problem: Data Silo = Data is not unified = Hard to know customers quickly
- 23. What is Support Data Platform? → All data is unified inside of TreasureData
- 24. Support Usage for Sales / Product / Dev team in Chartio
- 25. Enriched User Profile in Zendesk
- 26. Support requires a development skill :) embulk-output-zendesk_users with embulk & digdag
- 27. Improved a process for Customer Slack Channel
- 28. Do you like Slack?
- 30. For a smooth communication between Sales/SE team and customers, but... Problem - Difficult to do a tagging for categorizing a question. - For better KPI, we'd like to add a tag into all inquiry. - Need to join Slack channel - We can't help customers if we didn't join the slack channel. - No SLA - If a customer has an emergency issue for affecting their business in holiday and ping us on slack, Nobody might knows the message. - Lost a message where come from - Slack is NOT todo app
- 31. BubbleIQ helps communications between Zendesk and Slack - We/Customer add :key: reaction - BubbleIQ Bot submits a ticket automatically - We can communicate between Zendesk ticket and Slack thread
- 32. Multiple Components = Huge Logs
- 33. Many Components = Many Logs
- 34. Difficulty in finding out failure jobs due to data silo/SQL Current Notification Flows: - Ex. Presto Job Failure due to an incident, let’s check affected jobs. 1. Extract job id/account id by using a failure message in presto log on master TD account 2. Extract executed user’s email by using job id in analytics TD account 3. Extract a contract information by using account id from salesforce info in analytics TD account for group by “Region” (JP? US customer?) 4. Write a JP/EN Notification and send it out via Mailchimp
- 35. IN Progress; Trying to “Search” instead of Query
- 36. Future Works
- 37. Future Works - Documentation / Localization - Various Customers/Senario - Data Engineer / Integrator / Digital Marketer - Digital Marketing? Gaming? EC? - Efficient Support Tool - For Log Analytics - For Workflow Analytics - Global Team’s collaboration - US / Japan / Korea - Sales / Customer Success / Solution Architect