Analytics

This week, I’ve assembled five favourite quotes that I use when discussing analytics.

“Not everything that can be counted counts, and not everything that counts can be counted.”
- Albert Einstein

There’s some debate whether Albert Einstein said this or whether it was only on a sign hanging on his office wall at Princeton. No doubt, he was referring to digital marketing! In our world, there’s no shortage of numbers. Instead, our problem is often having the right numbers. For me, this quote epitomises the challenges of creating good key performance indicators (KPIs). We can measure a lot these days. But just because we can measure something, it doesn’t mean that it’s important. Quite often the important things to measure in business, our KPIs, are hard to measure, and it’s often because we are thinking about measurement in the wrong way (see next quote). The challenge is to count the things that count (good KPIs) and to get the right numbers right (good data integrity).

“If your only tool is a hammer, every problem looks like a nail.”
- Abraham Maslow

For a long time, digital marketing analysts would try to solve every problem with data from a web analytics system. It’s getting a lot better, but there’s still a lot of it going on out there. We don’t need a single tool; we need a whole toolkit. Having good web analytics is absolutely necessary, but rarely sufficient. In addition to web analytics, we need good voice-of-the-customer intelligence, an understanding of the competitive situation, and an appreciation of the actual technical performance of the channel. This has implications for the skills base and training for analytics teams within organisations and also levels of investment required.

“The price of light is less than the cost of darkness.”
- Arthur C. Nielsen

Art Nielsen was a pioneer of modern marketing research. This quote characterises the challenges in getting organisations to spend on measurement and analytics as they often view it as a cost rather than an investment. These days investment in data collection technologies may not need to be that high. There are lots of free tools for web analytics, voice-of-the-customer research, site performance measurement, and so on. However, there is no point collecting data if you don’t have resources to analyse, interpret, and act on it. It’s not a question of whether you can afford to invest in these resources, as Art said, it’s a question of whether you can afford not to. This quote should be written in the footer of every business case for investment in analytical resources!

“Most people use statistics the way a drunkard uses a lamp post, more for support than illumination.”
- Mark Twain

This one is one of my favorites. The weekly report is the scourge of most organisations. It goes out every week, because it’s the weekly report. Most likely, this report was started by someone who is no longer in the organisation. Most likely, it doesn’t tell anybody anything that they need to know. Next week, don’t send out the weekly report and see what happens!

“After all is said and done, more needs to have been done than said.”
- Neil Mason

OK, I cheated here, but this is along the lines of “actions speak louder than words.” The whole point of measurement and analysis is that it leads to different outcomes. Analytics is about making better decisions and reducing risk in the business. There’s no point having good data and sophisticated analysis if the results are not acted upon, either because of the organisation’s culture or business processes. Successful use of analytics requires organisational agility to bring change and measure results. That’s what we mean by optimisation.

So those are my favorite quotes about analytics. Has anyone got any others?

This article was originally published at ClickZ

Try searching online for an image of a dashboard and I guarantee that most of the images will have a gauge on them, and indeed most Excel plugins have gauge options. But does a gauge visualisation really warrant this level of attention? So, “To gauge or not to gauge?”. Not in my opinion unless there is a very, very good reason!

There is nothing wrong with gauges, but they take up a lot of space given the amount of data they impart. Generally, they are rather cumbersome and against the principles of pixel density or non data ink. I would argue that a gauge should only be used once on any visualisation and ONLY if it is the critical metric! Think of a car dashboard, and you may think of a series of dials, but the main visualisation is for speed – a rather critical metric when driving. So if a metric is important it should be front and centre, but I would also argue that there is a better visualisation – Bullet Charts

The chart on the left shows the same information as the first gauge, but takes up much less horizontal space – space which could be used for additional insight! The grab on the right shows how bullet charts could be used in a horizontal orientation to show performance of a range of metrics against target values. Here we are showing the performance of a metric (shown as a blue bar) against a target, with a: red(ish) backgound for 0-75% of target; grey backgound for 76-100% of target; and a green background for above 100% of target.

As you can hopefully see, having a set of charts laid out in this ways makes it very easy to see the overall picture of the data and as such the performance of the business.

There are, of course, examples of gauges that cram in additional data, but they key question should always be – “Does this add value?”

What are your thoughts? Do you use gauges? Are bullet charts ‘sexy’ enough?

“The world has gone social; it’s just that we don’t know how to measure the value of that yet.” That was one of the messages coming out of the Adobe Omniture Summit here in London. Thirteen hundred people from 500 companies from 30 countries came together last week for two days to learn, share, network, have fun, and digest the views of various speakers on the issues of the day. The other message was that the world’s going mobile – “the eyeballs are moving from fixed web to mobile web” – and there’s loads of challenges to measuring that too.

From the keynotes from the likes of Christian Hernandez of Facebook and Brian Solis at Altimeter, the emphasis is that businesses and organisations need to focus on people and to understand the social consumer. Hernandez talked about “putting people at the centre” and the power of understanding how enabling social interaction can bring disruption to different industry sectors. First of all, there was photo sharing, then there was gaming, next in Facebook’s sights is commerce; social commerce. According to Hernandez, the classical shopping funnel is dead and the shopping funnel is in fact a circle from awareness to interest, to decision, to action, to recommendation, and back to awareness. If that’s true, then it’s going to play hell with those pretty funnel reports in your web analytics tool. But Hernandez said that the future of social marketing was all around measurement and data.

However, Solis reminded us that “we cannot measure what it is that we do not know how to value” and therein lies the rub. What is the value of a social media marketing strategy? Solis’ advice was to make it actionable. Social marketing is about actions, reactions, and transactions. If it’s about actions and transactions, then it’s more measurable. But at the heart of it Solis tells us that organisations need to understand the socially-connected consumer. What is it that makes them tick? What do they need or what is it that they want from you? To do this, it’s going to be important to start to segment out your social customers from everyone else and look for those differences in patterns of behavior and attitudes from everyone else. The trick is also not to treat “social” as one big lump but to also look at the different channels within social media such as Facebook vs. Twitter, because, as some of the case studies shown demonstrated, there can be interesting patterns between what people do when using the different social channels. In fact, different types of people might be using different channels to do different things, so you can’t take a one-size-fits-all approach. Segmentation is key.

Once you understand the social customer, then Solis says to introduce a social marketing program that works for both of you, and to do that you have to think like a connected customer and give them something to talk about. Often what the social customer wants (i.e., a special offer) is different from what the organisation wants (i.e., a relationship), so you’ve got to find those things that are going to work for both of you. The key thing though is to make sure that you create outcomes; a “click to action” as Solis calls it. If you have outcomes, then at least you have something to measure against.

Of course, I know that these things are easier said than done. Organisations have been struggling for years to define “what good looks like” for their web channel strategies, let alone their social channel strategies, but defining the expected value is the critical starting point to building a social media measurement strategy. By defining the value, you’ll be able to better understand the metrics that will tell you whether that value is being obtained, and therefore the technologies can give you those metrics.

This article was orignially published by Clickz

There has been a lot in the news recently about privacy. I previously wrote about problems Apple were facing regarding the perception of user tracking. Wednesday’s headline in ‘The Independent’ newspaper relates use of social media to circumvent super-injunctions and a Scottish paper has just published a full front page picture of an individual claiming to have taken out a super-injunction.

Privacy is a hot topic. Do the ‘public’ have a right to know?

The issue of privacy has been a concern for a long time, with companies often focusing on the use of customer data to maximise their profits rather than their customer’s experience. This short sighted view will lead to a back lash from customers, unless they feel in control. Nobody likes to receive spam emails containing completely irrelevant content, but at the same time, people feel uncomfortable about being tracked.

The upcoming EU cookie legislation takes effect on the 25th May, requiring companies to request users permission before setting any cookies not deemed to be “strictly necessary” for site functions. Whilst it is unlikely that we’ll see any real impact from this legislation in the short-term, it does show an interesting trend to great user empowerment. Our own research shows that most sectors are poorly equipped to deal with this legislation and they will have to act soon, to avoid being the first pulled over the hot coals.

A recent article over at Mashable highlights the efforts of ‘Buyosphere‘ to help customers regain control of their data. The company essentially allows users to choose which brands can have access to their data, allowing for a win-win relationship. Again, this shows the continuing trend of customer empowerment.

The next few months should be extremely interesting, as the issue of privacy is at the forefront of the news. Ultimately, it is the end consumer that should be the winner in the privacy argument. Companies are now focusing their efforts on engendering trust and convincing their customers to share their information with them. But in a world where even the giants of the technology, such as Sony, can lose customer data it is vital that people know what they are agreeing to.

With power comes responsibility!

For more on Foviance’s Data Privacy Audit

I’m an analytical and technology geek, and so of course I think so. But, perhaps a better question is should a chart be sexy?

For me, the central purpose of a chart is to convey insight. It must be meaningful, else what’s the point? There is the classic criticism of 3D Pie charts that were the mainstay of almost all Excel charts for a long time. Nearly everybody that I know used the 3D feature of Excel when charting data. The reason? It looks cool (in a post 1990 sort of a way!) The slight downside is that it is very difficult to see the actual proportions of each segment when presented in 3D due to the skew caused by simulated perspective.

2D Pie charts are better are conveying precision, and from a purest standpoint, a much better visualisation. However, in my experience it is also important to ensure that there is an emotional engagement for the visualisation. The aesthetic is important too, as it allows you to capture the attention of the viewer. There is no point in having an insightful chart if no one wants to look at it!

Over the years of creating dashboards at Foviance, this balance between form and function has been an interesting journey. We rightly have constant debates on the subject, both internally and with our clients. There are often conflicting views based on the business requirements, the people involved and their backgrounds. For some, Stephen Few (function) is king and for others Edward Tufte (form). The concepts from Stephen Few, such as pixel density, are extremely important and act as a great rule of thumb. However, it is also vital that the end user can relate to data on an aesthetic and emotional level, and here the concepts from Edward Tufte can be great guide.

Many dashboards and visualisations try to be the perfect solution, but the reality is that the ‘one-size fits all’ approach is fundamentally flawed. Everyone is different and everyone processes information differently. It is vital that this be taken into consideration when designing a visualisation to ensure that the meaning can be conveyed as efficiently as possible. For example, we have recently built a dashboard which is intended to be used by both executive users as well as analysts. It has therefore been necessary to build not just a flat dashboard, but a dashboard application that will over just the top-line metrics for the executive users (along with a commentary), but that also allows analysts to drill-down into each of the metrics for additional context and insight.

Of course, it is not always possible to build an all-singing/all-dancing dashboard. We have budget and time constraints, and so we are forced to compromise the form and function across the various users. Is this a bad thing? No, it forces us to be innovative in our thinking and to focus on ROI. But it should be in the back of everyone’s mind when displaying data: Who will be viewing this? Why are they viewing this? What do they need?

At Foviance, we have moved away from static score cards, to more dynamic dashboards allowing the user to interrogate the data, to become more engaged, and to get more insight. The next step will be to offer a far more personal experience whereby the same dashboard could be viewed in different ways by different users. A simple example, may be to present different colour combinations for different users, or to allow the user to display a data table rather than a pie-chart.

Data is important, but insight is vital.

Over the next few weeks, I’ll be posting a few specific examples and it would be great to hear what you think.

So, are you a 2D or 3D person? Can a chart be sexy?

We all dread the weekly report. A spreadsheet arrives in your inbox, crammed full of numbers, perhaps a few charts, maybe even some attempts at ‘data visualisation’ but usually without any narrative or explanation. You look at it and think “So what?” So what is it telling you and so what are you going to do about it? We’re not short of numbers. We’re short of understanding about what to do about the numbers. As analysts, we’re complicit in this. We have a habit of sending out data rather than insights. We’re good at reading numbers but not great at telling stories. We need to be telling more stories because people remember stories and they rarely remember a piece of analysis.

But what are the ingredients of a good story and how is that relevant to analysis?

Stories have a beginning, a middle, and an end. When it comes to delivering results of a project in a presentation, I urge consultants to go by the old adage: “Tell them what you’re going to tell them, tell them, and then tell them what you told them.” In other words, start off by giving key highlights, tell the story, summarise, and close.

In addition to a basic structure, a story needs a good narrative. In our data-driven world it’s easy to get hung up on numbers and to have slide after slide containing tables full of numbers and complex graphs. The storytelling analyst will understand the narrative and will find a way to tell the narrative in an engaging and memorable way. Often, less can be more. Fewer numbers can give greater insight if they are the right numbers. Less precision can lead to more confidence. “67.3% of visitors…” can almost appear to be too precise and can leads to challenges around data accuracy, whereas “two thirds of visitors…” takes the data issues away and the focus is then on what the two-thirds of visitors did or didn’t do or think.

The storytelling analyst will also develop the flow. Stories have a cause-and-effect relationship. There are events and then there are consequences. Otherwise, the narrative is just a series of events and there is nothing for the listener or the reader to take away. In the book ‘Elements of Persuasion‘, Richard Maxwell and Robert Dickman define the elements of a good story as:

• The passion with which the story is told
• A hero to drive the action
• An obstacle or an antagonist to challenge the hero
• A moment of awareness where the hero realises how he can overcome the obstacle
• A transformation in the hero and the world around him

It can be difficult at first take to see how these elements might be incorporated into the delivery of a piece of business analysis or research, but it’s worth thinking about it. Passion is possibly the easiest in a way. A good analyst will be interested in the business and will have empathy for it. He will understand the consequences of the events that he’s describing and the relevance of them to the audience.

The story’s hero could be business customers or a segment of them. One technique might be even to personalise the story around an individual or groups of individuals by using pen portraits or personas. This can help bring the story to life and increase its relevance. The obstacle that the hero needs to overcome would be the point of the research or the analysis. The moment of awareness would be the insight from the research and the transformation would be the consequences from the recommendations accompanying the analysis.

So, next time you send out the weekly report or deliver analysis, ask yourself “What’s the story here?”

This article was originally published by ClickZ

Apple has been in the media wars of late, with some significantly negative publicity relating to the tracking of users via their iPhone. Whilst the reality is rather different to that initially portrayed, it does highlight some interesting trends…

For those of you that watch South Park, a recent episode highlighted the public perception of end-user licensing and privacy statements. The character Kyle doesn’t bother to read the privacy agreement of an Apple iTunes update and has to suffer the consequences, which cannot be detailed here!

Apple has suffered a significant damage to its public reputation recently, due to the news that it has been tracking the location of it’s iPhone and iPad users.  Apple, along with Google, have been requested to participate in a US Senate panel on the issue of mobile tracking, with Apple adamantly stating that it does not track user’s locations, but uses the data collected to help speed up the use of location based services. As you can see from the map below (produced using the iPhone Tracker tool ), the wi-fi location and cell towers around your phone, can give an impression of your own location – you’d never guess that I use the east coast railway a lot!

<Click to enlarge>

Last year I wrote about the issues of advertising on mobile devices, a theme that I’ll be coming back to again shortly, along with the potential difficulties of mobile tracking. One of the issues related to the latency of the tracking and the corresponding issue of data accuracy. A recent article by ‘Localytics’ has highlighted this problem and it will be an ever more important issue as the use of mobile devices becomes more and more main stream. The common consensus (Wikimedia & Online Marketing Trends)  predicts that mobile devices will surpass desktops as the method of browsing the internet by 2014.

The Apple story is interesting as the issue of tracking on mobile devices, and indeed this specific issue relating to the user’s location, has been known about for well over a year. It only gained momentum in the popular press when the information was visualised!

This really shows the importance of visualising data. We all love a good story and need to be able to relate to the protagonists. Data is great, but insight is better!

Our own Neil Mason discusses the use of storytelling in his recent blog post. We are often confronted with the conflict of making data look ‘sexy’, think of glossy 3D pie charts, rather than meaningful. For me it is vital that a data visualisation tells a story. It should allow the user to experience a journey through the data to gain meaningful insight.

What is interesting, is that the depiction of data can often feel like an uphill struggle, everyone has an opinion on what a chart should look like. It is often only when you have played with a couple of different visualisations that the true meaning comes through. From a linguistic or memory perspective, this would be termed as ‘semantic’ meaning that a greater level had been achieved. As analysts, it is our job to help create this journey for our client.

This is not easy. It may take several iterations before the insights come through, but then again, it wouldn’t be a true story without a struggling hero now would it!

For more about Foviance’s Data Privacy Audit

As MVT becomes increasingly important, Econsultancy have just released their Multivariate Testing Buyers Guide. The report provides tips and advice for identifying the right MVT vendor, including detailed trends and issues affecting the online testing sector. John D’Arcy, Analytics and Insights Practice Director at Foviance explains:

Foviance have worked with a number of the MVT vendors in this buyers guide and as with any data technology there are tools that fit business requirements better than others. We find that once technology issues are solved, businesses still have challenges in achieving value from optimisation tools. When they get the technology and added value combination right, the benefits can be enormous. Where else can you get these sorts of immediate uplifts?

A big challenge is delivering a set of tests that will impact your business immediately and offer meaningful learning. To do this, you need to take a step back and deliver a prioritised roadmap of tests before you start out. Understanding the potential impact of a test can be analysed with web analytics data. However you get much more actionable information by learning from user experience tests, using tools to listen to the voice of your customer and from competitor benchmarks. Prioritising your tests based on these data sources will ultimatley deliver much more value. We find clients who are able to do this coherently are able to deliver a test roadmap that provides both immediate commercial potential and long term customer experience impact.

For more on Foviance’s Multivariate testing and Analytics Services.