Recommended Reading – 9th September 2010

The final group of links I think are worthy of your attention are below

• Paul Graham provides his perspective on Yahoo!’s problems. In part, he thinks they should have continued to think of themselves as a technology company and not a media company

• Dino Demopoulos has created an elegant perceptual map of location based services in terms of people/places and now/then

• Neil Gains has some good tips on how to frame questionnaires – clear questions provide clear answers

• Joel Rubinson has some good ideas in his post on brand building for the digital age

• Meghan Keane writes on e-consultancy about behavioural targeting, and how it needs to strike the right balance between usefulness and creepiness

• JP Rangaswami has a typical thought-provoking post on the nature of expertise and knowledge asymmetries in an online environment

• And finally, this Oatmeal comic succinctly portrays the reasons why working from home is both awesome and horrible

sk

Advertisement

MRS Mobile Insights Conference

The first Mobile Insights Conference, hosted by the Market Research Society, took place in London this week. I was in attendance.

I’ve mentioned my thoughts on the word insight several times in the past. Did I take away many insights from the day? No. Were the talks useful? In parts. Was it worth attending? Yes

Although the first use of a mobile phone for research was over ten years ago, the techniques available on the devices are still in their infancy. This means we are still in the experimentation stage of understanding what can be achieved on mobile – compelling innovations and applications of research do exist, but are rare.

As such, the projects discussed over the course of the day were largely (but not exclusively) small and tactical. It would have been great to hear of any companies really embracing mobile research to inform strategy, but perhaps we haven’t yet got to this stage.

Given the relative lack of application, the day largely focused upon the mechanics of mobile research. A topic not particularly attractive to research buyers – I counted two research buyers on the delegate list alongside attendees from research agencies and suppliers (discounting speakers). While it important to collaborate as an industry, the realities of competitive advantage meant that, with the odd exception, research agencies aren’t going to be particularly candid about learnings when they are speaking to other research agencies.

Those gripes aside, I did take some several useful things from the day.

Tim Snaith, OnePoint Mobile Surveys – Defining the Future

After some opening remarks from this OnePoint colleague Neil Jessop, Tim Snaith took the floor to talk about the immediate future. Some of the points he made include:

• “If we define mobile research in a certain way, it will be adopted in a certain way”. Instead, we should be open-minded as how people ultimately choose to engage with the mobile will define how we can engage with them
• There is currently no knowledge base in mobile so we need to build foundations with case studies and open reports
• Downtime on the computer is decreasing because there is so much on there to keep us busy. Mobile is now the device for downtime, creating opportunities for surveys
• Blending computer and mobile research isn’t ideal, since the space restrictions on mobile mean you are effectively asking different questions
• International mobile research needs to consider the different regulatory and technological requirements in each market (for instance, different networks use different character sets in SMS)

Alex Wilde, Globalpark – Tracking trends

Alex defined the four types of research that can be conducted on a mobile as a survey (e.g. invited by a banner ad), diary, storytelling (capturing stimulus) and netnography.

An issue with surveys is that there are over 3,000 combinations of mobile handset, browser, operating system and so on. Ensuring usability is therefore a challenge. Within survey design, he advocated drop down boxes rather than radio buttons, and limiting text boxes to a single line field

With regards to some of their mobile studies

• They receive 35% of responses within the first hour (Lightspeed say they received 60% of completes in the first 15 minutes and 90% within the first hour – evidently research design will be more indicative than overall methodology trends)
• When survey respondents were asked after a few questions if they would like to continue the survey and, if so, where; 70% of iPhone users said they would continue on their phone but 90% of non iPhone users said they would prefer to continue on their computer
• 90% of respondents were willing to share their location via GPS

Tom Webber, Nielsen – Smartphone trends and opportunities

Tom shared several statistics with the audience:

• Smartphone penetration in the UK grew from 12% to 20% in Q1 2010 (smartphone defined by an operating system such as iOS or Android)
• As the prices of phones rises, long term ARPU falls (presumably this was down to unlimited data charges and the need to bundle SMS rather than charge individually?)
• Net Promoter Scores for smartphone owners are fifteen points higher than those for featurephone owners (26 vs 11)
• 64% of iPhone users use apps daily, compared to 52% of Android users and 38% of other phone users
• The Apple App Store has 84% satisfaction; the Android Marketplace has 81%. Apple performs better on range of apps and download experience; Android wins on ease of browsing and discovery
• American 13-17 year olds sent an average of 3142 texts a month in Q1 2010 (I believe this figure includes twitter and facebook updates); the figure for under 12s was 1152
• In the US, Twitter accounts for 42% of phone messages sent; Facebook for 16% (though I believe GSMA data has Facebook far ahead in the UK, at least in terms of time spent)

Patrick Hourihan, Yahoo! – APPetite

Patrick (Disclosure: an ex Essential employee although, like in Ghostbusters, our streams never crossed) was the first speaker to give a presentation, rather than read a report, and it was probably my favourite talk of the day in terms of interesting content (if not practical application).

After some general industry statistics (such as 11m social network users via mobile, 5.7m people downloading games and 11m using mobile media for entertainment, he ran through the methodology for APPetite

• Online diary and forum
• Focus groups in London and Manchester
• Depth interviews with in situ mobile usage
• 2,000 person online survey among 16-65 year old mobile media users (done online as mobile doesn’t give the depth or the same level of representativeness)
• Mobile app survey using Research Now’s real-time data collection tool (I think the questions were regarding which app was used, where and how it compared to online)

Things the research covered either qualitatively or quantitatively include

• Consumer expectations have shifted from functional use to emotional use
• Nearly as many use mobile for entertainment (58%, if you include social networking) as they do functional (63%)
• There is a claimed decline in snacking, with longer usage sessions
• Brands have been mapped to times of day, with the BBC, Facebook, Google and Yahoo! in use across the entire day
• Apps have the wow factor among users, and satisfaction of services (such as Facebook) is higher for the app than the for the web (NB: This wasn’t broken down by operating system)
• However, there is some confusion over what an app is (e.g. it could be a  shortcut to a website)
• 55% don’t have a preference between an app and web so long as they can fulfil their specific needs
• When a mobile service doesn’t work, 44% blame the brand, 34% the phone and 29% the network

Siamack Salari, Everydaylives – Using mobile for ethnography

Siamack’s talk was the only brazen sales pitch of the day, but it was justified since he was the only person speaking with a genuinely groundbreaking (or so it seemed to me) research technique – a mobile application where people can capture audio, video and pictures, tag it and upload it to a central server to be sorted, filtered, analysed and edited.They can comment and feedback on the information others compile.

The iPhone app cost £5,000 to develop (whereas the BlackBerry app cost around £60,000) and will allow ethnography to be conducted on a larger scale. Previously, he would film 2 hours of footage a day over several days, conduct interviews separately and then spend time editing the reams of footage.

The app come with a small price, but he is shortly going to launch a consumer version, which is free and would allow people to effectively sell their lives to willing buyers. He mentioned that, independent of this, Coca Cola and P&G approached him about fitting out their workforces with the application.

Learnings from his first four studies are

• Not all iPhone users are familiar with iTunes, and so will need to be walked through app installation
• A template of how to shoot video is required for a consistent quality of submissions
• Research needs to be designed carefully to avoid it altering people’s behaviour
• He has facilitated post-filming opt-outs, where he gives people an email address and a code, in case they want to have their privacy protected in their friends’ film

Leonie Hodge, Channel 4 and Anthony Cox, Sparkler – Using SMS to uncover drivers of viewing

As a cost effective alternative to ethnography (and allowing a greater sample), Channel 4 commissioned Sparkler to conduct research into drivers of viewing using 24 depth interviews, 9 in-home evening visits and 35 media diaries lasting for 6 days using a combination of mobile and paper.

Essentially, on days 1, 3 and 5 Sparkler would text the respondents to ask them what they were going to watch on TV that evening. They would send several messages (manually) if requiring clarification or reminding. On days 2, 4 and 6 people would fill in a standard consumption diary of what they watched the previous evening.

The research showed that there were very few appointment to view programmes (a disappointment?), and the majority of consumption was last-minute or spontaneous decision making.

Sparkler’s advise was to

• Get the stakeholders to pilot the technique, in order to engage them with the process
• Keep it as simple as possible for the respondent
• Design the outputs so analysis can be as efficient as possible

AJ Johnson, Ipsos MORI – Using mobile to gain greater customer insights

AJ was the exception in that he was quite open and candid about Ipsos MORI’s various experiments with mobile research – both successes and failures. Initiatives include

• Using mobile for passive media consumption for radio (via ambient sound recording) and posters (via GPS)
• Quick polls for PR purposes
• Comparisons to online research, which showed that mobile surveys took longer with shorter answers and a lower response rate, but that quality of response (recall, in this instance) was higher
• Multimedia diaries – they invited 1,000 people to take part in recording their weekend activities. 200 agreed to take part, but only 37 provided usable material. They had asked for 4 diaries a day from people, but received an average of 2.7 per person (they tended to be photos)
• Software they have been experimenting with includes Microsoft Pivot and The Link (unsurprisingly, I can’t find the link)

Gavin Sugden, T-Mobile – Measuring customer satisfaction via SMS

Gavin was also exceptional, in the sense that his was the only example of SMS research being used more strategically. T-Mobile changed their customer satisfaction survey from a 20 minute CATI survey to an SMS survey of around 6 questions – 3 standard questions and 3 from a battery of 15. They found SMS would be a better capture method and more cost effective than CATI, IVR or the mobile web, though Gavin didn’t seem to have a problem with privacy, saying that people could delete the messages if they didn’t want to respond (T-Mobile customers would be contacted within 24 hours of purchase in-store, web activity or phoning customer services).

With people only eligible to complete one survey in 60 days, they are seeing a 20% response rate.

They have an online reporting tool, which allows analysis of quantitative data and aut0-coding of verbatims.

They currently use the results for performance management (e.g. benchmarking stores) but are moving towards coaching and development of staff, and seeking of ways to improve customer satisfaction. They are also going to start include campaign awareness questions

Liam Corcoran, Fly Research – Insights into alcohol consumption

Liam walked us through a case study from a study he ran for Mintel, looking at alcohol consumption of young adults. Rather than hazy recollections, they used mobile research for in situ response. At 10pm each night, they would send an SMS with a link to a 7 question mobile internet survey.

He also talked a bit more generally about their mobile panel research.

• They tend to receive 60% response rates from their panels and 20% from cold sample
• 50% of their responses come within the hour and the vast majority within 24 hours
• He wouldn’t recommend a mobile internet survey of more than 15 questions
• Mobile is a great tool for a follow-up to online research, if just a couple of additional answers are needed

Sarah Sanderson and Jason Vir, Kantar Media – Integrating mobile with other research techniques

Sarah and Jason gave the last talk by going through a very recent project they ran during England’s World Cup campaign (presumably the project was ran specifically for the Conference; it must be nice to have the resources to be able to do that!)

The research was designed to access the emotional response in real-time, which they felt would be more honest and truthful

If I remember correctly, they:

• Recruited people to the project based on answers to questions in TGI Sport+ and TGI Postscript (the TGI omnibus)
• Got some people to take part in an online community to discuss the tournament
• Asked these people to submit videos and pictures of them during/after the game (I think 7 people did this, I’m not sure if the community was larger)
• Sent a WAP survey at half time to a larger sample. 207 responded, 90% of these during the half time interval

Neil Jessop closed the day with a few remarks, including a prediction that there would be more mobile surveys than online surveys. This was the one contentious comment of the day (in my opinion), and this sums up my overall feelings of the event. There were a few interesting asides and thoughts, but there was nothing at the event that really made me stop and think. A shame, but perhaps I was expecting too much.

sk

Image credit: Research-Live

Scheduling online video is a good move

TV Week reports that online video providers are increasingly looking to their TV counterparts for best practice. This is a sensible approach – while the two platforms are different (and should be treated as such) there are plenty of lessons to be learned.

Find that advertisers wait for audiences before committing isn’t a major discovery, but one of the moves announced is very savvy.

Regularly scheduling transmissions is a very good move.

Benefits of online video include the asynchronous nature, letting people choose what to watch when and where they please. But unless people are aware of what is available at any given time, how will they know to consume it?

Video consumption may be impulsive (as Tania mentions in the comments to my previous post), but that impulse is prompted. It will either be driven by an external notification such as a comment or an advert, or a memory.

I’ve found that people tend to visit TV-related sites with specific content in mind. Their visit may not have been planned, but their journey is determined by the content they know is available. Few visit to generally browse – going back to a point in my previous post; it is content driven. People’s first visit to Hulu will be prompted by word of mouth, a search or a publicised piece of content (e.g. Tina Feylin). They then become aware of the variety of content on offer (living in the UK, I am very jealous), and this memory gives another reason for subsequent visits.

TV, to an extent, follows the same path. Channel-flicking is more prevalent but people do plan in advance to watch specific shows. Push factors include

• TV Guide – people looking at a schedule to see what is on at a point where they plan to view
• Rituals – they know their favourite shows or strands are on at the same time on a given day
• Conversation – personal, word of mouth recommendation is a major factor
• Adverts/Trailers/Bumpers – giving a teaser of what to expect

TV has that live, event feel to it that online video doesn’t. The mass viewing with watercooler chat is an important element of TV. Online doesn’t scale as well, but among niche groups (or tribes, as the buzz word du jour would have it) there may well be a hunger to immediately consume and deconstruct new content. And the opportunity should be provided.

Therefore, as well as a regular schedule, prompts should also be incorporated. These could be email reminders, an RSS feed, text updates or an online TV guide that incorporates web series/online video on demand. Forums may be asynchronous but they do operate to the rules of time and as proved by many commenters, there is that urge to be “first”.

This is definitely a move in the right direction. From an advertiser perspective, will it make it easier to plan? Partially, since the timings of availability are known. But there will still be the long tail of consumption well after this, as people find out about the show and catch up. (Though people will presumably watch the show in order, and so the campaign can still be progressed over time in the intended manner.)

The key is the call to action. A schedule isn’t enough – a specific prompt is vital to driving consumption. We see it with TV driving people online, and this will be no different.

sk

SIDENOTE: I see that Yahoo!’s record video views were driven by a combination of topical and viral videos. This is unlikely to be consistent as topical videos are event driven and virals are inherently unpredictable (see my work in progress typology), but it means that when a perfect storm hits, the spikes will be dramatic. Congratulations to them for a great month.

Image credit: http://www.flickr.com/photos/smb_flickr/