Data-Driven PR Metrics: Share of Voice, Competitive Benchmarking, Correlations
3 likes4,557 views
The document presents a framework for measuring public relations effectiveness through data-driven methodologies, focusing on share of voice, competitive benchmarking, and correlation analysis. Key concepts include measuring performance against competitors, the importance of quantifying metrics, and using sentiment analysis to refine insights. The insights are aimed at improving public relations strategies by identifying industry trends and best practices.
Data-Driven PR Metrics: Share of Voice, Competitive Benchmarking, Correlations
- 1. DATA-‐DRIVEN PR MEASUREMENT Sandra Fathi President Affect @sandrafathi web: affect.com blog: techaffect.com email: sfathi@affect.com eMetrics Summit San Francisco, March 29-‐Apr 2, 2015 Slides: www.slideshare.net/sfathi
- 2. ABOUT ME • Sandra Fathi • President, Affect • Public RelaJons, Social Media, MarkeJng • Council of PR Firms • PRSA Past PosiJons: – Tri-‐State Chair – NY Chapter President – Technology SecJon Chair 2
- 3. ABOUT AFFECT 3 Technology Healthcare Professional Services:
- 4. PR MEASUREMENT Three Concepts for Discussions: • Share of Voice • CompeJJve Benchmarking • CorrelaJons @sandrafathi 4
- 5. PART I: SHARE OF VOICE
- 6. DEFINITION Share of Voice: Comparing your crucial performance metrics against those of compeJtors or the market. • You have to measure something • What you measure needs to be analyzed proporJonately against compeJtor data (or market data) to establish market share @sandrafathi 6
- 7. THE FORMULA Number of ConversaJons Including Your Company = X * 100 = % SOV Total ConversaJons on a Topic @sandrafathi 7
- 8. ADVERTISING CONCEPT 25% SOV 75% SOV @sandrafathi 8
- 9. SHARE OF VOICE I 72% 28 % ConversaJons Talk About Me @sandrafathi 9
- 10. SHARE OF VOICE II 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Q1 Q2 Q3 Q4 CompeJtor C CompeJtor B CompeJtor A Our Company @sandrafathi 10
- 11. SHARE OF VOICE III 300 ArJcles MenJon My Company 145 ArJcles MenJon CompeJtor 589 Industry ArJcles 87 ArJcles MenJon Both 51% SOV in the Industry @sandrafathi 11
- 12. KEEP IN MIND • Share of voice should be defined for a period of Jme (finite start and end). • Share of voice is oien most useful when limited to a single plajorm or medium. For example, business press coverage or Twiler. • Share of voice can be overwhelming if trying to look at too large a segment or industry. Try choosing SOV among top compeJtors or in key interest areas. @sandrafathi 12
- 13. SOV: SOCIAL MEDIA ANALYTICS PLATFORMS @sandrafathi 13
- 14. SOCIAL MENTION @sandrafathi 14
- 15. SIMPLE EXCEL FORMULA @sandrafathi 15
- 16. ONLY PART OF THE STORY • Doesn’t consider senJment • Doesn’t consider sources (exclude self produced/owned media) • Doesn’t consider quality, only quanJty (Is NYT blog same as obscure geek’s tweet?) • Don’t accept the data blindly – human verificaJon is required with any tool @sandrafathi 16
- 17. OTHER APPLICATIONS & CONSIDERATIONS ConsideraJons: • Apply senJment or tonal filters (posiJve/negaJve) • Apply qualitaJve measures (by Jer or by type) ApplicaJons: • Industry trends/hot topics (i.e. SOV on cloud security) • Specific products or services • Broken down by geographic or demographic parameters (i.e. SOV in 18-‐25 market) @sandrafathi 17
- 18. PART II: COMPETITIVE BENCHMARKING
- 19. DEFINITION CompeJJve Benchmarking: The conJnuous pracJce of comparing a company’s pracJces and performance metrics against the most successful compeJtors in the industry. • You measure processes and results • You must idenJfy a ‘benchmark’ or indicator that will be a unit of measure to compare • The desired outcome is to understand which processes lead to greater success (best pracJces) in order to improve your company’s performance @sandrafathi 19
- 20. COMPETITIVE BENCHMARKING • IdenJfy my compeJJve set for comparison • Choose my units of measure: press coverage • Set parameters: top 20 business and trade • Define a Jme period: 6 months • Choose a tool (news monitoring service) or begin manual research @sandrafathi 20
- 21. EXAMPLE: RADWARE ObjecJve: • Build & Maintain Radware’s PosiJon as a Thought Leader on ADC and Security • Maximize Radware’s Overall Public RelaJons Results Strategy: • Compare and Contrast Radware’s Press Release Output with Top 3 ADC and Security CompeJtors • Compare and Contrast Radware’s Coverage with Top 3 ADC and Security CompeJtors • Analyze Results • Apply Best PracJces and Lessons Learned to Radware to Improve Overall Performance @sandrafathi 21
- 22. EXAMPLE: RADWARE ApplicaJon Delivery Network Security @sandrafathi 22
- 23. • Analysis of press release strategy and resulJng coverage over 6 month period • Specifically as it relates to relevant products or business units • Only in top 20 business and industry/sector publicaJons METHODOLOGY @sandrafathi 23
- 24. METHODOLOGY II @sandrafathi 24
- 25. RADWARE PRESS RELEASES Security 43% ADC 27% Both* 12% Other* 18% Press Releases * ‘Both’ includes releases related to both security and ADC, ‘Other’ includes non-‐product releases (e.g. company news, financial announcements etc.) Press Releases Security 14 ADC 9 Both 2 Other 6 @sandrafathi 25
- 26. ADC COMPETITORS: SECURITY & ADC * ‘Both’ includes releases related to both Security and ADC 11 11 2 15 46 39 31 92 0 10 20 30 40 50 60 70 80 90 100 Radware A10 Citrix F5 Press Releases ArJcles PRESS RELEASES VS. NUMBER OF ARTICLES @sandrafathi 26
- 27. 9 11 2 7 44 16 31 38 0 5 10 15 20 25 30 35 40 45 50 Radware A10 Citrix F5 Press Releases ArJcles ADC COMPETITORS: ADC ONLY PRESS RELEASES VS. NUMBER OF ARTICLES @sandrafathi 27
- 28. ADC COVERAGE BY TYPE 0 20 40 60 80 Radware A10 Citrix F5 Other Report Commentary AcquisiJon Partner Customer Product Product Customer Partner AcquisiJon Commentary Report Other Radware 7 2 1 0 1 34 1 A10 1 1 10 0 1 0 26 Citrix 27 0 0 0 2 0 2 F5 24 0 3 0 12 0 37 @sandrafathi 28
- 29. COVERAGE QUALITY 0 10 20 30 40 50 60 70 80 90 100 Radware A10 Citrix F5 74% 10% 23% 50% 26% 90% 77% 50% MenJons Features FEATURE VS. MENTION @sandrafathi 29
- 30. ADC CONCULSIONS • Number of press releases did not correlate to number of arJcles • Radware was leading in SOV on key topic (ADC) amongst compeJtors and the quality of coverage by comparison was significant (ValidaJon!) • Overwhelming majority of Radware’s ADC coverage was generated by reports (ValidaJon!) with product and customer news trailing far behind • CompeJtors were leading with product news and capturing media alenJon (Opportunity!) • No one was successfully telling the customer story (Opportunity!) @sandrafathi 30
- 31. SECURITY COMPETITORS 14 23 28 10 84 164 68 68 0 20 40 60 80 100 120 140 160 180 Radware Arbor Imperva Prolexic Press Releases ArJcles PRESS RELEASES VS. NUMBER OF ARTICLES @sandrafathi 31
- 32. SECURITY COVERAGE BY TYPE 0 50 100 150 200 Radware Arbor Imperva Prolexic Other Report Commentary Alack AcquisiJon Partner Customer Product Customer Partner AcquisiJon Alack Commentary Report Other Radware 8 9 6 0 28 22 11 2 Arbor 19 1 1 18 29 44 52 0 Imperva 2 1 1 0 8 18 19 19 Prolexic 0 0 0 0 43 2 14 9 @sandrafathi 32
- 33. COVERAGE BY QUALITY 0 20 40 60 80 100 120 140 160 180 Radware Arbor Imperva Prolexic MenJons Features 35% 65% 31% 69% 34% 66% 54% 46% FEATURE VS. MENTION @sandrafathi 33
- 34. SECURITY CONCLUSIONS • Radware is #2 in overall SOV but the quality is not as strong (more menJons vs. features) • Leading customer and partner conversaJons (ValidaJon) • Good job at Story Hijacking (responding to security hacks) but room for improvement (ValidaJon) • CompeJtors winning at report coverage and commentary (Opportunity!) @sandrafathi 34
- 35. CONSIDERATIONS • Good for understanding what worked but not necessarily ‘how’ it worked • Costs for research may outweigh benefits of insights • Once you’ve idenJfied the ‘best pracJces’ you may or may not be able to replicate them • Consider non-‐compeJtor companies to benchmark • Do you want to ‘emulate’ or ‘innovate’? @sandrafathi 35
- 37. DEFINITION CorrelaJon: A mutual relaJonship, or interdependence, between two or more things. • In the absence of being able to prove ‘causality’ you may be able to demonstrate a ‘correlaJon’ to demonstrate the impact of a PR or markeJng program • A correlaJon is posiJve when the values increase together • A correlaJon is negaJve when the values decrease together @sandrafathi 37
- 38. TYPES OF CORRELATION Source: MathisFun.com 38
- 39. THE FORMULA 39 Pearson’s CorrelaJon: @sandrafathi
- 40. FUNCTION IN EXCEL 40 @sandrafathi
- 41. CORRELATION IN EXCEL 41 @sandrafathi
- 42. FUNCTION IN EXCEL 42 @sandrafathi
- 43. SCATTER CHART 43 @sandrafathi
- 44. LINE CHART 44 @sandrafathi AcquisiJon
- 45. 3-‐D LINE CHART 45 @sandrafathi
- 46. MULTIPLE DATA SETS 46 @sandrafathi 0 500 1000 1500 2000 2500 3000 3500 4000 4500 Q1 Q2 Q3 Q4 Sales Web Traffic Press Coverage
- 47. SPURIOUS CORRELATION 47 @sandrafathi Source: TylerVigen.com
- 48. 48 @sandrafathi Source: TylerVigen.com SPURIOUS CORRELATION
- 49. CONSIDERATIONS • User correlaJons cauJously and don’t trust the math blindly • The visuals oien tell a story as well • Remember that correlaJon is not causality, it can only help as an indicator or potenJally predict probability • Data is sJll beler that your opinion 49 @sandrafathi
- 50. FINAL THOUGHTS • In measurement, speak the language of the C-‐Suite • Excel is sJll the best dashboard for data visualizaJon • Don’t be afraid to learn that you are wrong • Don’t be afraid to change direcJon • Use the data to gain execuJve support – Strategy – Resources – Headcount – Budget 50 @sandrafathi
- 51. THANK YOU CONTACT: Sandra Fathi President Affect @sandrafathi web: affect.com blog: techaffect.com email: sfathi@affect.com Slides: www.slideshare.net/sfathi