Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

Webinar | Good Guys vs. Bad Data: How to Be a Data Quality Hero

Duplicate contact and account records. Missing field values. Inaccurate or outdated information. Unstandardized data. Typos upon typos. If you manage and operate sales and marketing tools, you’ve likely encountered these bad data scenarios (and others!).

We get it — data quality isn’t the sexiest topic. But the impact of poor data quality is undeniable, causing 21% of marketing budgets to be wasted, according to research from Forrester and Marketing Evolution. Furthermore, factors like increasing competition and evolving buyer needs continue to make data health even more important. Improving data quality can stretch your marketing dollars further, enable operational efficiency, and act as a strategic growth lever.

Tim Liu (Head of Product at Hull) and Brad Smith (Co-founder & CEO at Sonar) have spent their entire careers working in data integration and operations, so they’ve seen it all. In this webinar, they’ll share:

- Data quality nightmares they’ve personally dealt with
- Common scenarios where bad data can rear its ugly head
- Proactive strategies for getting ahead

  • Be the first to comment

  • Be the first to like this

Webinar | Good Guys vs. Bad Data: How to Be a Data Quality Hero

  1. 1. Thank you for joining! The live broadcast will begin shortly to accommodate last minute attendees. 1
  2. 2. Before We Begin ● Don’t be shy! ● Got questions? Use the questions module! ○ We’ll address them throughout the presentation. ○ If we have time at the end, we’ll run through a few more. ● Webinar replay will be sent to all registrants within 24 hours. 2
  3. 3. 3 Brad is the CEO and co-founder of Sonar, a change management platform designed to make things easier by improving process, communication, and documentation for cross-functional teams. Brad is also the founder of Wizards of Ops (or WizOps for short), a community Slack channel where ops people can share tips, ask for advice, and make operations magic. GET EXCITED Featured Speakers Tim has spent his career building and scaling software products from the ground up. The last decade he has worked primarily on data startups in a number of different verticals. Brad Smith CEO, Sonar Tim Liu Head of Product, Hull
  4. 4. 4 How is everyone today? POLL TIME!
  5. 5. What’s the biggest obstacle keeping you from clean data? 5 POLL
  6. 6. WhatWe’llCover 6 What do we mean by “bad data”? The operational and financial impact of bad data Prevention and mitigation strategies 1 2 3
  7. 7. What bad data scenarios have you faced? 7 POLL
  8. 8. Incomplete contact and account profiles Unstandardized data Duplicate data records Typos What Do We Mean by “Bad Data?” 8 Legacy fields (that have not been properly sunsetted) Data flows gone wrong Merging records incorrectly
  9. 9. Downstream Impact of Bad Data 9 Inaccurate or suboptimal targeting Missed sales opportunities Loss of customers Reduced productivity Poor customer experience Inaccurate performance reporting “Death by a thousand cuts”
  10. 10. 21%of marketing budgets are wasted as a result of bad data. — Forrester and Marketing Evolution, 2019 10
  11. 11. When bad data strikes, what do you feel is the biggest misalignment? 11 POLL
  12. 12. ● Merging or expanding departments ○ e.g. M&A, organizational restructuring ● Executing too quickly on growth initiatives ○ e.g. with new funding rounds ● Onboarding new tech and data sources ● Status quo isn’t working anymore When Bad Data Strikes 12 Usually due to a misalignment of: PEOPLE PROCESSES TECH
  13. 13. Proven Strategies 13 1 Begin with the end in mind 2 Gain visibility 3 Architect &execute 4 Measure & maintain
  14. 14. Begin with the end in mind. 1 14
  15. 15. ● Don’t try to boil the ocean; start small ● Determine your ideal outcomes ● Beware of scope creep ● Acknowledge unknown complexity and try to minimize risks upfront Recommendation #1: Plan ahead to mitigate risk. 15
  16. 16. Gain visibility. 2 16
  17. 17. ● Customer data platforms are designed to unify and centralize ● Identify duplicate data ● Identify gaps in the data (i.e. What fields are missing data?) ● Where is the “best” data? Which sources do you trust more? Recommendation #1: Put all of your data into one place. 17
  18. 18. 18 ● Understand how your metadata is being mapped ● Which system is updating which field? ● How do you know who owns what? Recommendation #2: Create a blueprint.
  19. 19. Architect and execute. 3 19
  20. 20. ● What is identity resolution? ○ The process of taking the data points that define an entity from different online or offline systems and merging them to create a single, consolidated, and consistent record of that individual or company. ● How will you define your entities (e.g. individuals and companies)? ● Several strategies: ○ Natural keys, universal IDs, multiple natural keys Consideration #1: The identity resolution strategy. 20
  21. 21. Consideration #2: The data model. ● The data model defines how entities relate to one another ○ i.e. Leads and Contacts associated with Accounts ● What is the “source of truth” for these relationships? ○ How are these relationships created? ○ How are these relationships updated? ○ What are the controls to surface issues and fix them? 21
  22. 22. ● Data transformation tools, like the Hull Processor, can help ○ e.g. update all instances of ‘USA’ to ‘United States’ ● What is the Hull Processor? ○ A real-time editing environment for your customer data, the Hull Processor lets you compute new attributes from existing data and update customer records on the fly. Consideration #3: Data cleansing and standardization. 22
  23. 23. ● Training and user enablement ● Documentation is key ● Defining and sticking to your naming conventions Consideration #4: Promote consistency. 23
  24. 24. ● They’re inevitable! ● Backup your data, if possible ○ Especially if you’re doing large scale cleaning operations ○ Any way to “Undo” is your friend ● Always provide an escape hatch for manual correction ○ Engineers love to automate, but if there’s no possibility to manually intervene, you’ll have to rely on that engineer to fix issues Consideration #5: Edge cases. 24
  25. 25. Measure and maintain. 4 25
  26. 26. ● Reminder: Begin with the end in mind ○ What goals did we set at the beginning to make sure our project was a success? ○ Define what success looks like for your company, paying attention to holistic revenue goals. ○ Be mindful about setting aside vanity metrics. ○ Know what data points you need to be collecting in order to gauge momentum and progress towards your goals. Begin with the end in mind Measuring Success 26
  27. 27. ● Nothing is ever static ● Tactical things to put in place: ○ Alerts and notifications ○ Audit trails and commit messages ○ Project owners ○ Sandbox organizations Managing a Changing Data Landscape 27
  28. 28. 28
  29. 29. THANK YOU! Visit >>> Join >>> Visit >>> 29