How Does Mobile Compare?


Published on

Presentation by Pala Kuppusamy and Steve von Bevern
(presented at MRS Mobile Research 2013)

Published in: Technology, Business
  • Be the first to comment

  • Be the first to like this

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

How Does Mobile Compare?

  1. 1. 2Smartphones Now Dominate Mobile Share InMost Developed Countries• 2013 estimates say 1.4 billion smartphones and 268 million tablets inactive use (not just units shipped)• Android leads in smartphones; is in tablets• EOY 2013 estimates:• 45 million Windows phones; 20 million Blackberries(3.2% and 1.4%, respectively)• Android = 798 million handsets (57% share); is = 294 million devices(21% share)
  2. 2. 3
  3. 3. 4Geo Surveys
  4. 4. 5Behavioral
  5. 5. 6Rich Media Uploads
  6. 6. 7In-The-Moment Surveys
  7. 7. 8
  8. 8. 9Home EthnographyHow comfortable are respondents are withqualitative aspects of mobile research?Research Now Mobile invited 3,500 panelists to takepart in a home ethnography study designed tounderstand respondents‟ willingness to participate inqualitative parts of mobile studies, particularlymedia uploads.CASESTUDY• 679 completes in 2 days,despite the study being long &media intensive• 97% scanned barcodes• 99% uploaded pictures• 82% uploaded audio• 80% of those with petsuploaded videos
  9. 9. 10Sporting Event DiaryHow do researchers collect in-the-moment data atthe 2012 Olympics?Research Now Mobile recruited for 40 respondents tocomplete two two-day diaries for people at the 2012Summer Olympics or watching them at bars orat home.CASESTUDY• 69 respondents acrossboth diaries• 62% uploaded images• 19% uploaded video• Impressed with detailedopen-ended responses• High level of respondentengagement
  10. 10. 11Holistic InsightsHow does a digital newspaper use passive datacapture to gain deeper reader insights?A digital newspaper that caters to tablet usersthrough a newspaper app uses Research NowMobile+ to send targeted surveys and overlays thesurvey data with behavioral data fordeeper insights.CASESTUDY• Behavior data like appdownloads, music played,etc. are passively captured.• Survey data is collectedperiodically by pushingsurveys.• Data from repeat surveysandpassive data capture are tiedtogether for deeper insights.
  11. 11. 12Retail SurveysHow does a major CPG company get shopperfeedback on their in-store displays?Research Now Mobile created a mobile survey forshoppers, whose locations were validated by GPSdata, to complete while inside targeted supermarkets.CASESTUDY• Respondents uploaded photos ofin-store displays and scanned UPCbarcodes of products• Geo-fencing ensured the survey wasavailable only when respondents werein the store, ensuring strongvalidation and data quality• Rich In-The-Moment data and insights
  12. 12. 13
  13. 13. 14How Do Mobile Panelists Differ from Online?Comparisons (Un-Weighted):U.S. PanelMobile Online Gen Pop CensusGender Male 51.1% 42.3% 49.2%Female 48.9% 57.7% 50.8%Ethnicity Caucasian/White 84.1% 88.5% 72.4%African American/Black 5.5% 2.7% 12.6%Hispanic 7.2% 3.2% 16.3%
  14. 14. 16How Do Respondents Feel AboutMobile Surveys Vs. Other Modalities?What is your satisfaction with this survey?Respondents who complete a survey on mobiles are more likely to rate theexperience as „somewhat‟ or „very fun‟ compared to those completing a moretraditional online survey (13% vs. 68%).Online respondents are more likely to be „neutral‟ about the entertainment valueof the survey (65% vs. 26%).0%10%20%30%40%50%60%70%Boring SomewhatboringNeutral SomewhatfunVery funOnlineMobileTotal
  15. 15. 17How Do Respondents Feel aboutMobile Surveys Vs. Other Modalities?The entertainment value of the survey is more highly rated across almost all theage groups in our Scout sample, compared to the online sample, which is moreclosely aligned to mobile ratings.0%10%20%30%40%50%60%70%18-2425-3435-4445-5455-6465+Fun by Age and Sample OriginScout Sample Email Sample Online Combined Mobile Sample
  16. 16. 18Are There Device Preferences?Device preference is very heavily skewed by age, with younger age groups from18 -44 years preferring a mobile device. Over 45 years of age, the preference forcomputer completion increased exponentially as respondent age increases.0%10%20%30%40%50%60%70%80%90%100%18-24 25-34 35-44 45-54 55-64 65+Mobile DeviceComputer
  17. 17. 19Are There Device Preferences?Mobile sample respondents overwhelmingly prefer mobile devices for completingsurveys; the age effect is demonstrated. In the oldest group, 25% of those whocompleted a survey on a mobile device preferred to complete a study on acomputer.0%10%20%30%40%50%60%70%80%90%100%18-24 25-34 35-44 45-54 55-64 65+Mobile SampleMobile DeviceComputer
  18. 18. 20Open Ended Response DifferencesUS Mobile Base %# of respondentswith an answer86 100 86%Avgas (non-zero mean)length18Those with1-2 word response2 100 2%Those with5+ word responses68 100 68%Those with10+ word responses43 100 43%US Online Base %# of respondentswith an answer656 663 99%Avgas (non-zero mean) length 8Those with1-2 word response139 663 21%Those with5+ word responses391 663 59%Those with10+ word responses190 663 29%Mobile open ends show consistently richer results
  19. 19. 21Open Ended Response DifferencesUK Mobile Base %# of respondentswith an answer84 100 84%Avgas (mean) length 17Those with1-2 word response5 100 5%Those with5+ word responses65 100 65%Those with10+ word responses50 100 50%UK Online Base %# of respondentswith an answer861 868 99%Avgas (mean) length 9Those with1-2 word response146 868 17%Those with5+ word responses541 868 62%Those with10+ word responses309 868 36%
  20. 20. 22
  21. 21. 23DesignCollecting the Survey Data: Sampling• Replicate invitations / quota plans• Five days in field• 6 Minute Median LOI / 89% QR / <3% abandon
  22. 22. 24ComparisonComparing the Survey Data (Weighted to Demos)“What I do at work #MarchMadness RT:@espn:Yes. It is still totally acceptable to fill outbrackets all day -”“Its official...#MarchMadness is back! Howmany of you plan on using @CBSSports"boss button" to hide your habit at work?”
  23. 23. 25Do Respondents Answer Differently WhenTaking Mobile Surveys?Generally, respondents do not have issues completing a survey; only a tinysegment find mobile surveys „somewhat‟ (5%) or „very difficult‟ (1%)to complete.0%10%20%30%40%50%60%70%80%Very difficult Somewhat difficult Neutral Somewhat easy Very easyOnlineMobile
  24. 24. 26How Do Respondents Feel aboutMobile Surveys Vs. Other Modalities?While there is very little difference in the ease of use rating between online andmobile respondents, there is a decrease in rating the ease of use score amongstthe older respondents.0%10%20%30%40%50%60%70%80%90%100%18-24 25-34 35-44 45-54 55-64 65+Overall EaseOverall Difficulty
  25. 25. 27Conclusions
  26. 26. 28
  27. 27. 29Best Practices for Mobile Research• Sample selection/recruitment• Target your respondents…natural response rates will skew younger, higher income• Use both online and mobile to insure broader representation• Questionnaire Design• Don‟t be fooled: shorter questionnaires are not a „rule of thumb‟ for mobile• Use a journal approach if you require more• multiple surveys (dairies) work well• Avoid conjoint if it requires real time „pings‟ of another server• ≠ flash media in native app tools• Use of rich media• Keep use of audio, photo or video capture to <5 requests per survey• Enhance open ends….consider audio instead of text responses• Diary studies, qualitative, secret shopper• Mobile is great for diary & secret shopper work.• 5 to 10 day diaries generate good response rates• Invites—geo fence, target, venue• Use of reminder invites…best practices
  28. 28. 30Are There ‘Downsides’ to Mobile Research?• Geo locations are great, but plan wisely• Do not plan quick „drive bys‟ without larger over-samples• Geo invites require lat/long for locations targeted• Tradeoffs of battery life vs. accuracy of measurement• Avoid conjoint studies or others that require real-time „pings‟ withother servers• Adaptive conjoint, max diff, etc.• Survey Length• We see wide length variation.• Don‟t shy away from longer surveys.• < 10 minutes is recommended or longer with good responses but longer is possible• Multiple point of engagement on mobile over course of time.• Consider diaries• multiple surveys fused together into one set of results (store visit, in-home use, etc.)• Native app is best confined to high share devices like is or Android devices• Blackberry, other small share devices: limit them to web app today
  29. 29. 31Are There ‘Downsides’ to Mobile Research?• Be prepared to deal with rich media• Showing rich media or capturing rich media…• …Rich media generates a lot of data• …Be prepared to code, tabulate and analyze audio, video and photos• Diverse platforms mean you may never achieve 100% coverage• Use “LCD complexity”: confine it to single choice and multi choice if want all devices• Richness of survey vs. device coverage• Cover Android and is; monitor Windows and Blackberry
  30. 30. 32
  31. 31. 33