Archive for the ‘Social media’ Category
Andrew Bruce Smith: proving social media ROI with Google Analytics
March 27th, 2013
Andrew Bruce Smith, Director, Escherman
Google Analytics is ubiquitous – 86% of British businesses have installed it, even if they’re not all using it. It’s free to install – but the true cost is in understanding how to get the most out of it. And it can prove the direct and indirect value of social media.
When you login, you are confronted with a default screen. Social is probably buried in the referral traffic. When you look at it, most people assume that search and PPC drive the most traffic. To get the most out of Google Analytics’ attribution tracking, you need to define some goals. ANY event that happens on a website can be a goal. Few people really define the purpose of a site. Do that, set goals and assign values to them.
Social media has a marriage problem – the priest gets 100% of the credit for the marriage under the last click model. What about the guy who introduced them? Or the restaurant where he proposed? These are the factors you’re looking to draw out.
Those companies that do this see that social is an assistive medium, not a closing one. It’s often an early part of a valuable pattern. One ecommerce site got the majority of its traffic from search – but social delivered people with a higher propensity to buy. If they’d turned off social, their sales would have fallen off a cliff.
We can move away from focusing on top line numbers, and start couching the business case for investing in social media, because you can track both direct and indirect value from social activity.
People rarely take your brand value at face value – that’s when they turn to social. That’s why it’s an assistive medium. People don’t just click on a link and think “I’ll buy that” – but attribution analysis shows that downgrading investment in it is unwise.
Panel: Is quality the forgotten metric?
March 27th, 2013
Luke Brynley-Jones leads a panel debate, on the idea that quality has been forgotten as a social metric. What do people think?
Sharon Flaherty, head of content & PR, Confused.com: How do you know if anything’s quality content? I get e-mails from people all the time asking me to host their infographic with an embedded link, for SEO purposes. How do I know if it’s quality? Have they done their research? Guest blogging is becoming dangerous, because it’s all about SEO.
Lutz Finger, co-founder of Fisheye Analytics: You need to act like a journalist. People used to hail that social networks got way from the gatekeepers. That means that people need to act like journalists to verify the quality of content. There is a need for it. It’s an interesting cycle backwards.
Robin Grant, managing director, We Are Social: What defines a quality fan base? Raw numbers don’t cut it. Engagement is arguably a better number – because surely only quality content would attract engagement? Well, possibly not. Are pictures of tomato ketchup – which get great engagement – actually having any impact on the brand? It’s almost impossible to track that. You can have facile, lightweight engagement.
Katie Howell, MD, Immediate Future: It’s about layers. There are a whole series of layers – but I totally disagree that you need a journalist. There are different skills needed. People need to understand their audience in detail, and what the purpose of the content is. Most journalists aren’t trained in that. If you get into optimisation, you figure out how to add another action to engagement. The rhetoric is that social media is a conversation. With one of our B2B clients, the best performance is where there’s no conversation.
Sharon: We got into content seriously when the Mayday update pushed us to page 2 of Google. It worked, but it took two years. We use Google Analytics, and have it in the place that they can track people’s journeys from the content they land on through the site.
Lutz: What we measure needs to have an aim. If you don’t have that, you can’t build measurement. Quality is “quality for a certain aim”. If I’m trying to sell my book, I’m giving speeches that get people to sign up for information on my book. If they sign up, it was good content. LinkedIn endorsements immediately triggered companies offering endorsement bots. What makes an important person? My wife and kids are important to me…
Robin: “What’s the worth of a Facebook an” may be a tedious question, but it’s important and can be answered. For Bulmer’s we figured out it was worth about £200 per annum. There’s a lot you can do to understand the value of segments of your audience.
Sharon: You should start from the point that everything you produce is the best you could produce. Then it’s just a question of how you slice and dice it. Our no claims bonus guide is boring to me – but its a huge driver of traffic and conversion. Is that quality? Yes.
Katie: Platforms are forcing us down certain paths when it comes to content creation.
Robin: Relevance is user-centric, and applies to any form of engagement. Quality is something that drives something meaningful for the brand. It’s a subset of relevant.
Sharon: We all understand relevance, and we all think that. t’s about making sure you have relevant content, but with quality at the same time.
Lutz: You need relevant content to trigger actions. Quality is subjective, relevancy is the one link throughout.
Lutz Finger: the rise of the influential social media bot
March 27th, 2013
Lutz Finger, Co-Founder, Fisheye Analytics
There are fake social media profile out there. Why? What’s their business model? That’s what I want to discuss.
7% of Tweets are fake. This are not spam – these are friend. 20% of us accept friend requests from people we don’t know and check out. Once these guys are in and talk to you – they’re good. They have personality. This army of bots is with $2.76m. Can an army of bots be the new influencers.
The first generation of bots were just spammers going social. They’re cheap and easy to create. The post messages very fast. They’re basically shouting. But their conversion rate is 1:1.25 million. That’s a terrible conversion rate. They’re also easy to spot. They have few friends, they over-use hashtags, they post all the time, or in bursts, and they are lonely. Their only friends are other bots. Twitter has acted and spam dropped from 10% of a Twitter traffic to 2% between Feb 2009 and Feb 2010.
But by 2011 only 20% of fake Facebook accounts are being detected. Twitter is suing spammers. But the business mddl is changing – bots are being used for smear campaigns. Many politicians have had sudden spikes in followers, that are clearly fake. These weren’t bought by the person in question – they were assigned by their opponents, looking to make them look desperate.
Spammer are learning, too. So bots 2.0 are learning to be social. They aim to be your friends and then to build trust – and through that, influence. Could they create mass movements? Social media allows things to become contagious. It’s easy to use, and people can “support” things with a click. But can people be influenced? The idea of the influential person isn’t true. Humankind is so stupid as to follow just a single person.
Single data points – like Kimmel talking about the double rainbow video – don’t disprove this. They can’t repeat that phenomenon.
People’s willingness to be influences is topic dependent, for example, people influence each other on classical music, but not indeie or alt music, research shows. Influence is often homophilia in reality. Book and hotel reviews are used by 22% of people – but 40% of those reviews are fake. We like to hear things from multiple sources – 50% of people get their news online -= but 55% of journalists rely on social media for stories, and aren’t great at verifying…
We’re in an arms race, and we need to ensure we don’t train bots anymore. OKCupid is a data-heavy dating service. The more thy shut down bots, the more they train them. So they don’t do that any more – they just move the bots into a secondary world, just full of bots.
Marshall Sponder: making your measurement and analytics tools actually work
March 27th, 2013
Marshall Sponder, Analyst and Metrics Consultant
Ultraviolet data is the data that’s all around us, but which we don’t see and thus don’t use. How can we capture this data, and is it worth it? The data we need is often the data we already have.
In the 1950s and 60s we had big computers and structured data. In the 1990s the personal computer led to a shift towards unstructured data. How much do you want to look into that big pot of data? How much value does it have to you? I have so much unstructured data now, I have to decide (as a company) what I’m going to look at, and what questions I need to ask myself before I look at it. We’ll soon be living in a work where lifestyle choices make you a less desirable person, because you can use the data to figure this out.
In the social space, for every new channel there are now forms of analytics. Your level of investment in platforms in this is going to be defined by your business needs and business size. You need data, and you need to understand the business problem it could solve – and then you pay for the workflow. As your sophistication increases, so too will your investment.
Most marketing implementations are a mess – a mass of platforms that don’t talk well to each other. The biggest challenge is hooking up motoring, social CRM and engagement tools so that the information starts making sense. Things like Google Glass are just going to create more and more data, which means more and more platforms will emerge to analyse that data… All the platforms are trying to do everything – but they don’t all do everything equally well, so you might have to buy a couple of them to get what you need. USe cases dictate the features, but because everyone is chasing the same customers, they’re all claiming to have all features. The danger is that you end up using two systems that do the same thing.
Somemo - a framework that looks at listening platforms from the standpoint of the user. Two thirds of people buying these platforms are internal or external agencies. People are looking at the wrong reasons: the metrics and the interface. All the vendors are targeting the 6.6% who spend more than $10,000 a month on tools. Over 50% want to spend under $100 a month…
Nearly 66% of people end up unhappy with their platform…
You need to understand where you’re at, to understand what you can do with data. Do that, and you’ll make better choices…
Panel: Which social media metrics matter?
March 27th, 2013
Conferences are full of award-winning social media campaigns, but no chairman is starting a company report celebrating a fourfold increase in Retweets – how can we connect the coal face with the boardroom, asks Philip Sheldrake of Euler Partners, chairman of the panel.
Andrew Bruce Smith, Escherman: Obviously, the best metrics are the ones that most match our business goals. Olivier Blanchard is a big advocate of ROI as pure cash. Our industry is wallowing in non-cash metrics, so no wonder we’re struggling to make the business case.
Jacqui Taylor: For me, it’s about context. What do the massive numbers Obama’s team pout out about the economy actually mean? That’s why we create blended teams. Metrics evolve on from KPIs – but it’s certainly not Likes. Completely disagrees on ROI. Social data’s actually the key – but nobody will believe us for five years.
Matt Owen, eConsultancy: Money is what matters to the board. Those of us working on this are still looking at last click attribution. Likes and follows are vanity metrics – you can buy them – it’s all about targeting.
Vanity Measures and the attraction of big numbers
Sharon from confused.com: Is using promoted tweets vanity?
Matt: We used them for about six months. A lot of them are really badly targeted. If you have the time and inclination, you can target really well. $15,000 went in and $70,000 came out.
Andrew: A very large multi-national brand created a Facebook page and got 60,000 Likes in a few months – and the UK division was told to do the same. The UK agency had a strong suspicion that those Likes had been bought. The reaction? “Find me the cheapest place to buy Likes…” The board demands big numbers, so all they wanted was to provide them.
Katie: What if all the readers of a newspaper called you at once? What if all your followers do? If you can’t give a good business reason for big numbers, why should you do it?
Jacqui: Everybody who is involved in social is everyone in the company. It’s not just about social media specialists. Just look at the leaks from the company, look at everyone’s social media accounts. It’s a layer that goes across the organising. People are recognising that the power in the company is the people. We find hot spots and cold spots across the company, and socialise the best practice from the hot spots. The best evangelists you’ve got are your people.
Philip: Social media in most cases is still just lipstick on a pig.
Matt: Most companies still see social just as a marketing function. I think marketing is a function of social media.
Andrew: PR has always been under-invested, because it’s so hard to prove attribution. 90% of media consumption is done through a screen. Print is maybe 4%. Much social mica starts on a phone – we’re in a social and mobile world.
Katie: Here’s the thing: PR used to be about improving your relationship with stakeholders, which was difficult to monetize. But now you can monitor the health of relationships. You can watch the relationship with your CRM, or some other tool. IF you forget the attribution point, you can measure contribution to the health of relationships. You can divide relationship metrics, and apply it to social media.
Jacqui: Social isn’t what you do, it’s who you are. There are strands to it, like there are in your DNA. You need to rebuild your company around it. This is not companies who hire based on Klout score. It’s about companies who create their businesses around social data. The groundwork still needs to be done – connecting resources is pretty hard even with board-level support.
Social Media ROI
Philip: Is ROI just a question people asked when they are scared? Why do C-suite people never ask for the ROI of mowing the lawn?
Katie: Every single day there are social media people doing cost benefit analysis. There are financial acceptable ways of measuring this – what’s the cost of doing something with or without something – and the impact?
Matt: Any return on social media is just a base minimum, because it’s only the direct rests. You can’t track recommendation and word of mouth in the same way. A lot of our metrics are actually guides.
Jacqui: Very few decisions are evidence-based. One CEO asked how many people worked for his bank. It took six days, and he got four different numbers… This is an investment from a long-term point of view. We build a community. We’ll get more sophisticated over time, and start having real returns. But it needs to be a continual investment, and you can’t just dip in and out of the metrics.
Philip is calling for comments and feedback on Influence monitoring work on his blog
Case Study: PEER 1 hosting and social customer service
March 27th, 2013
Leon Chaddock, Sentiment Metrics
Leon quickly introduces his tool, which was the basis of PEER 1‘s social media monitoring work. The company was founded in 2005, started as blog monitoring and has expanded to social media monitoring. Often works through resellers, and concentrates on social customer experience and service helping businesses learn from customer conversations. They’re aiming to be a single, joined-up tool.
Maria Jose Serres, PEER 1
What does the biggest host in the world do when it knows that there are people talking about it on social media – some of them complaining, some of them potential leads? When they know the people they want to recruit are out there? They started listening – using sentiment metrics. But they needed a strategy, because without that they couldn’t measure. So they decided to connect the conversations they were finding with the right people in the company.
They used Sentiment Metrics to filter the conversations happening and target the ones they could actually act on. Complaints, for example, could be engaged with by directing information on service status to them. When people are angry, the first thing they do is tweet. The tool allowed them to find discussions buried in forums. They identified 55 sales leads with £800,000 in March. 10 complaints which could damage their reputation were dealt with.
Katy Howell: beyond the final click – true social media attribution
March 27th, 2013
Katy Howell, Managing Director, Immediate Future
We know that we need to start looking at attribution better, to figure out what actually influences our business. If we know that, we can allocate our budgets better.
There are two conflicting ways we can model generating leads from social media. If it was really straightforward, we’d be all bloody doing this now. Again and again we see people running programmes, and then ry to measure it afterwards. You need to plan in advance, and realise that social media is mainly about spreadsheets…
Shortlinks are simple and really easy to use – and give you the direct action. But, they’re last touch attribution, there are trust issues, and they don’t connect directly to conversions. Facebook recently launched conversion pixels, which is some hidden code for your website, which tags everything and creates conversion metrics for you. They can track across devices, too, which is unusual.
Last click metrics are used by most people – but it doesn’t tell you much. A lot of our ways of marketing have relied on this model. It assumes that there’s only a single touch point with your customer, and only looks at the bottom part of the sales funnel.
Nobody wins a chess game with just a bishop. It’s about all the players on the board – not just the one move that gets you to checkmate. You can get to this data – but not just via the last click. We need to be thinking about what role social plays in a multichannel attribution model. It’s really, really complicated. It’s a new way of thinking, but it’s creating some interesting insights – like looking at the first click is more useful.
Simplify to start with – look at your funnel and match anecdotal evidence with conversion figures. Social media is more a contributor -= 60% than an initiator – 30% – or a convertor – 10%, according to blended data from a number of online retailers. It sits around other elements of your marketing, because it’s so leaky.
But we don’t buy in this simple, linear way. She contemplated buying a Mac after years as a PC user – she talked to friends on Facebook, went to the Apple Store, talked to Dell on the phone, talked to her husband, went back to Facebook – and then searched for the Apple Store, and bought an iMac online. Last click showed that Google was the conversion point – but that ignores all the other contributors. If Facebook Graph Search takes off, we’ll see more people accessing smart.
This is not a very exact science – yet. Your assumptions can be wrong, and data can conflict because you’re not measuring something you should be. Other factors from rain to unemployment could change things for you. We need frameworks – because not doing anything is no longer acceptable. You should set rules and values – what elements need to be weighted more because they have more impact on your business. Are you testing hypotheses? If you put data together, you can see correlation – like Twitter activity driving search demand. Look for spikes and dips coming together.
Only 41% understand the interactions between digital and offline. It’s easy to over credit some channels – and attribution can be political. One size does not fit all – you need to customise for your circumstances. Avinash Kaushik suggested that your spend 10% of your budget on the tools, and 90% on the brains to use it. You need them to figure out what is really happening…
Jacqui Taylor: finding data in the social media explosion
March 27th, 2013
Jacqui Taylor, CEO, Flying Binary
We’re at a tipping point from a web and data point of view. We’ve been through Web 1.0, where companies were skeptical about website, through Web 2.0, where they were skeptical about social media. From a data point of view, we’re at Web 3.0, where social tools enable connections between people. The website is no longer enough. It needs to be blended with social media and other innovations. We’re still just at the beginning.
Gen Y (or Millennials) are an open, collaborative and most of all visual generation. Visualisations of information and data are here to stay.
Gen Z are the 1993 generation – the web generation. They’re just coming into the workforce now. They’re a kinaesthetic generation. They want self-controlled interactivity. They want to create their own stories from information. The Open Data roadmap for the government shows that most requests are from individuals – but it makes all the data available to anyone who wants it.
There’s a land grab in progress. Steer away from ROI – it’s return on influence in our world, if anything. The Arab Spring was actually a challenge to all governments around the world. Governments are being inning to understand this. OPen data work is allowing people to determine policy. The web is neutral – and it brings out who we are. People astroturfing need to understand that the web is permanent, and you’ll never be forgiven. Gen Y are particularly unforgiving. Astroturfing is short term gain, for eternal pain. Gen Z establish trust in a completely different way. They don’t establish trust any way apart from their own networks. They have moved away from command and control. Trust in governments is declining, but in media is growing.
Companies are starting to forget the notional firewall between their businesses and others, and are evolving into social businesses. How do you profit from this web? You blend social data into your business. The difference between internal comms and external comms is vanishing.
Over the next few years, another 4bn voices will join the web. That will change everything. In four years’ time we’ll have a global marketplace connecting 80% of the world. Is this a threat? Yes, to the command and control world. But social media is the key to the whole piece. The world is developing an ecosystem around you.
Big data? Fed up with the term. Open data is in danger of going the same way. Volume is the least important characteristic of data. It’s just data.
Social media specialists are comms specialists, and rarely love or are proficient in numbers. Web scientists are great with numbers, but aren’t great with communications. They create SWAT teams of mixed skills, with the social and numbers people directing operations.
If data is the new oil, social data is the new petroleum.
Jacqui Taylor, CEO, Flying Binary
Katie Paine: Standardising social media measurement
March 27th, 2013
Katie Paine, chief marketing officer, News Group
News Group is an international measurement and monitoring company.
She’s been working on standards for PR measurement for 15 years now. Most of the audience think we need standards for social media measurement – and she suggests three out of the six needed have been written. The aim is consistency, giving commonality of data that allows comparison. It’s not about one group or another. Initially IPR, amex and the global alliance collaborated on some ideas. She grew the initial group beyond PRs to include various more social media-centric groups. “Pretty much every letter in the alphabet was represented.”
Beyond that, they started arching out to clients, to get buying from them. So now, they had agencies, associations and clients involved – the three pillars were there. The path was clear: focus on the measurements that clients actually want.
The standard that emerged is voluntary, but uses a standard ISO process. You can find it on www.smmstandards.org
The first conclave happened in late 2011, and they had another one in New Hampshire last year. The proposals are posted for comment, and, after two months, published.
Priorities in order:
- Content sourcing and methods
- Reach and impressions
- Engagement
- Influence & relevance
- Opinion and advocacy
- Impact and Value
They came up with a content sourcing and methodology table – a transparency table. Everything from search strings used to how the sentiment analysis is done should be disclosed in the table. It shows what was calculated and how – and if that process was propriety.
Reach & impressions was taken on last October. The Digital Analytics Association led the work. ALL impression numbers are flawed – but multipliers should never be used, under this standard. Only 10% of Facebook is accessible to you, or Radian6. You don’t see 90%. Plus, EdgeRank means most of your Fans never see stuff. And then, if they could, they may not check their page often enough. Maybe a divider is more appropriate – 0.1% may? They prefer the phrase “opportunity to see” – and that’s specific to a channel. You can’t compare YouTube views to Facebook impressions directly.
ITEM = a post, micro-post, article appearing for the first time. Mention is a reference to the entirety being measured. Reach is the total mubre of people who had an opportunity to see the item. Impressions represents the gross number of items that could be seen be all people.
Engagement was next, and implies an action of some sort. Conversations are things that are about you – but are not necessarily engagements with you. This needs to be tied to your goals and objectives. Getting more likes is not your only objective. Remember 90% of engagement happens offline. Oh, and engagements can be positive or negative. Engagement should be measured by the percentage of your audience who engage by day/week/month. This can be down to item of content. This needs to be measured at the time or publication – a surge in comments six months later is probably spam. Engagement manifests different by channel, and should be measured differently.
Influence is something that happens beyond engagement. Give her another few days – and they’ll have a standard. Basically, influence means that your opinion or behaviour is changed in a way that would not have happened otherwise. It is not a single score or algorithm. It can be an outlet or individual, and it must be specific.
They’ve yet to address opinion & advocacy. Sentiment is, at best, 50% accurate. Impact & Value are always tied to your business objectives. The argument that Like is worth three bucks is completely specious. ROI is an accountant’s term – do not mess with it. Use cost effectiveness instead. Value can lie in both positive returns, and in avoidance of negative effects.
Implementation
The transparency table is being used.
Google+ shifts one gear up – everything changes
February 27th, 2013
Google+ has just shifted one gear up and everything will change in social media because of this acceleration.
Declan McCullagh, chief political correspondent for CNET, has just published a great story on the Google+ development, which effectively levels the playing field between my favourite social media sharing platform and the rest.
Agreed, it is puzzling why it has taken Google+ this long to make the move that allows third-party apps and websites to incorporate a sign-on service similar to Facebook Connect or Sign in with Twitter.
I think that Google got the timing right on this one. It sat back and watched carefully as Facebook and Twitter made their commercially-driven sign-on plays, noted the fall-out, picked through the debris and then delivered a better option.
Despite the compression of time that we all feel in the networked world, Google rarely rushes. I believe that is informed by the company’s guiding principle: “Do no evilâ€. Last time I looked, the company was not a saint and it has made mistakes, for sure.
Some of those mistakes might have generated evil. But the balance is substantially the other way, towards “socially good, relevant and usefulâ€.
It looks like Google+ has just released another social benefit on the net. It saw that Google+ streams could be overwhelmed if third-party apps or Web sites posted too frequently, through the sign on facility.
I think that it has found a way to balance users’ desires to share with the equally important desire by friends, family, and colleagues to be spam-free.
The CNET report quotes David Glazer, engineering director at Google+ as saying that the company is confident that its algorithms can find that balance.
“This is about taking relevance and applying it to social sharing. The principle is making the information available in the right place at the right time.â€
Google knows a thing or two about algorithms and I trust the company to reshape the social sharing economy.