Facts and Figures » Analytics

Social Media Metrics by Jim Sterne: A book review

Dan — Wed, 02 Jun 2010 06:30:00 +0000

Jim Sterne is one of the few elder statesmen of the web analytics world. Author, consultant, founder of the Emetrics summits and co-founder of the Web Analytics Association, Jim has been active in this field for over 25 years; his first book “Customer Service on Internet” was published in 1996.

Introduction

The importance of having Jim turn his attention to how to measure and optimize social media should not be overlooked. Here’s someone who has charted, chronicled and challenged many of the online developments we now take for granted. Whilst social media has been around a lot longer than we think, the “Social Media” that has got both marketers and CEOs both excited and anxious is still in its infancy. It’s the same social media that has now become headline-making and time-consuming, so that we are forced to say ‘hold on, are we getting anything out of all of this?’. This is where Jim steps in.

What I like immediately about this book (and I like books as my way of digesting information) is that Jim is very clear about what this book is and is not about. Thankfully, it’s not another diatribe about how game-changing social media is and you need to get your hands dirty with this range of tools. This book is about how to measure your investment – and although social media may be free, your time is not – in social media.

Who’s this book for?

In my view, these groups of people will find this book especially useful:

Organisations and individuals who have made some investments in social media and want to quantify the results they are getting
Those who remain sceptical about the chaotic nature of social media, but are afraid that they are about to get left behind – and therefore want to bring some structure to their future endeavours.
Practitioners and consultants who want to help their clients develop a system of measurement that will stand up to scrutiny, and can be discussed sensibly at board level, without hysteria or hyperbole

Sructure

I get a great deal of comfort from a book of this type that starts off each chapter with a verb. The author has understood that the reader wants to do something, learn a few things, take away a bit more and ultimately take some action. Jim’s favourite verb is “getting” – getting focused, getting attention, getting respect, getting emotional – see what I mean.

Chapter 1 starts as the book goes on – setting your objectives. No investment, social media or otherwise, can be justified unless it meets one of the three big goals, namely:-

Increase revenue
Lower costs
Improve customer satisfaction

The rest of the chapters, read chronologically, form a pattern of a traditional sales funnel. Chapter 2 is concerned with Getting Attention and reaching your audience, most importantly, not just any audience – but your audience. Chapter 3 – looks at how to spread your message through your networks in Getting Respect. Getting Emotional, the next chapter, recognizes that your message is an empty vessel unless we ascribe a sentiment to it. Understanding and classifying sentiment – positive, negative, neutral, sarcastic, humourous, etc. whilst not straightforward is vital in measuring the impact your social media activities are having.

Chapter 5 and Getting Response looks at the question of how you can track back the interactions you having the social media space to some desired outcomes. Getting the Message, in the next chapter, takes a small diversion from the ‘sale funnel’ approach and concentrates on the skills and techniques social media practitioners require for being both a good communicator, and most importantly, a good listener. Chapter 7, Getting Results, ties back your efforts to the overall objectives you set in Chapter 1. Chapter 8 outlines the way in different ways to ‘sell’ social media to the various tribes of doubters and naysayers.

The final chapter looks to the future and what the metathemes of social media might signify for the relationship between buyers and sellers in a networked world.

It ain’t new

Despite the furore and headlines around Twitter, it’s worth remembering that many forms of social media, in relative forms, have long histories. Message boards, review and opinion sites, blogging, bookmarking sites as well as media sharing platforms are well-known and familiar. The phenomenal growth in Twitter usage and Facebook’s rivalling of Google has naturally brought attention to how these newer community tools can be used. Some of this attention is directly from businesses and marketers working out how this form of ‘free marketing’ can be exploited to their benefit.

But a networked view of markets and buyers is not new either. The Cluetrain Manifesto of 1999, with its 95 theses, already understood that ‘markets are conversations’, and that most companies were failing to be involved in those conversations out of ‘obsolete notions of command and control’. From an offline perspective, David Ford and the Industrial Marketing and Purchasing Group, 9 years before Cluetrain, wrote persuasively about a chaotic, non-hierarchical view of industrial relationships. The Internet, and its newer forms, has simply given (?) these type of ‘relationship’ tools to the marketers.

Equivalence

One of the interesting themes that Sterne explores, is that of equivalence in the chapter, Getting Respect – Identifying Influence. Whilst the goal of many retention campaigns is to create ‘raving fans’, Sterne looks at one company’s approach to measuring fans. Baekdal of Baekdal.com defined fans as “one who either follows everything you do, or actively points other people to you – or both”. Based on his web stats, Baekdal was able to calculate that one fan was ‘equivalent’ to 445 regular online visitors. With an average fan having an average of 445 fans themselves, 15 fans commenting on a Facebook page broadcast this message to 6,675 others.

For those who have struggled to monetize ‘engagement’ activity, here was a clear-cut example. If you can put a value on a website visitor, and you can identify the equivalent number of visitors a fan is worth, you can monetize your social media marketing efforts. Added to that, if you can examine the success of fans in recruiting other fans, the ROI of social media just got a whole lot better.

Like every good book on web analytics there is the caveat about data purity, and Sterne makes the necessary right noises about the need for careful expectation setting. Data quality issues in the online space are well-known, and Sterne is right to make the case for correct presentation of the exactitude of the data, rather than sweeping the issue deftly aside.

Visualising and measuring networks

As well as putting a value on a fan, the book talks meaningfully about measuring ‘interconnectedness’ using a variety of tools. Companies like FMS Advanced Systems Group, Axiom, Rapleaf and Unbound Technologies have built tools that allow you to see, in visual form, how the form of these networks change over time and interact with one another. Whilst some tools focus on following ‘conversations’, others look at the linkages, via blogrolls for example, between actors. The goal of all this spidery network visualisation is to identify clusters, trending topics and important actors within your network. This could all be useful intelligence in targeting the right message to a significant group within your network.

Based on Pareto’s law, many of those in your network will be connected, but as Sterne asks, are they ‘respected’? Authority and impact are two measures for assessing the value of individual actors, whilst tools like Peterson’s Twitalyzer focus on generosity, signal velocity, clout and influence. However, like many activities in web analytics, you need to rely on more than one tool. Technorati may be great if many of your prospects and customers are blogging away; Twitalyzer is a powerful tool (and let’s not forget free) for those active on Twitter, but the real value, as well as hard work, will come when you come to aggregate these views into a single system. Until the technology catches up, it would appear you still need to do some pulling together of spreadsheets and Powerpoints for the exec team.

Sentiment

One of the reasons automation has struggled with social media metrics is that of recognizing sentiment. Whilst it is straightforward to capture mentions and retweets of items containing your brand name, what exactly were they saying about you? Think about emotions like sarcasm and irony, and see if you can work out ways for this type of emotions to be correctly classified. Not easy.

Although Sterne outlines frameworks for classifying sentiment – polarity and intensity – the text analysis tools necessary to automate and accurately classify millions of posts, tweets and reviews appears wanting, as Sterne opines

“Perhaps machines can eventually be taught. But who will teach them? Machines have a hard time with this because humans do, too”

Whilst we recognize that listening is just as important as participating in these ‘conversations’ their sheer quantity force us to look towards technology for the answer, when the real difficulty lies in codifying our own language. Layering on top of this semantic challenge is listening to those speaking in other languages, from different cultures and orthogonal perspectives.

One answer might be to look at what people do, rather than what they say – isn’t this after all the difference between market research and testing?

Measuring response

There are many ways to record actions of both prospects and customers, and Sterne has arranged this into a pyramidal “Engagement Food Chain”. The chain (why not a pyramid?) starts with Saw, advancing to Saved and Rated, and ends with Purchased and Recommended. Leaving aside some of the data challenges of identifying individuals, this model works well, particularly if you have an entirely online business model. This sales funnel view of the world makes the assumptions that the more people you have say, Commenting, ceteris paribus, the more you will have at the Purchasing stage. It’s classic stuff, and I am inclined to believe the underlying assumption.

Customer service and crowdsourcing

If you are still having difficulty convincing the Exec team after showing them how many visitors each fan is worth and how your volume of recommendations is positively correlated to your recent boost sales, try the Customer Service pitch. Many organisations have seen quick wins from listening to complaints from customers, and going on to help them out. For those, worried about getting involved in the ‘naked conversation’, using social media to resolve customer issues feels manageable, and can bring some personality. Virgin America, for example, has met those tweeting in-flight about poor service at the landing gate with “customer recovery coupons”. My sense is that, over time, customers will come to expect this type of interaction and woe betide the large corporations who have the wherewithal to do this, but sadly don’t.

One of the most interesting developments of social media is that of crowdsourcing – using the power of the many to develop innovative ideas as to how products and services can be improved. Asking your customer what could be done better, is not terribly new, but the volume of responses that can be garnered and the feedback loop as to what changes have been implemented make this a powerful tool. So when Dell.com gets 2,000 appeals for Linux to be pre-loaded on their laptops and PCs, it can respond, and did within three months of the first request. Measuring the number of ideas submitted, votes on each idea as well as further comments is a great way of understanding the power of your customer community. But of course, this is a big-brand exemplar, and you have to ask yourself: what would you think of Dell if it wasn’t doing something like this?

Return on investment

Overall, though, can we tie back our efforts to meeting business objectives? Jeremiah Owyang thinks so. He cites three critical measures to monitor in order to justify investment in social media:

Improvement in marketing efficiency (through greater amplification of your message);
Reduction in support costs (avoiding the costly call centre or a visit to the store);
Improvement to sales (subject to a robust tracking capability, of course).

With determination and thought, I am certain that most organisations will be able to use these three measures above to compare the investment they are making in social media with business outcomes

The future

Sterne at the beginning of the book promises not to be an evangelist for social media, and for the most part he keeps his word. But at the end of the book, the inevitable crystal ball-gazing chapter, John McKean of the Centre for Information Based Competition points out our central false assumptions. Our current thinking is based on the belief that innovation is occurring within the organisation.

In fact, McKean argues, “..the vast proportion of innovation is happening on the “buy side” of consumer interactions. The “sell” side innovations are progressively less and less.”. Organisations have failed to address their audiences in anything measuring like a success, when fail-to-respond rates are between 95%-98%. Instead of organisations continuing to try and ‘guess’ what the consumer wants, the efficient response is that consumers communicate their needs through a “consumer initiated dance (C2B) of Search, Find (engage) and Negotiate.” Like many codas to technology marketing books, only time will tell, but suffice to say when visits to Google are now being rivalled by those to Facebook, searchability is being replaced by socialability.

Evaluation

Jim Sterne has done an excellent job in pulling together many important strands from the body of, often self-inflated, research and practice of social media. He has stripped out much of the puffery in the debate, and concentrated on the fundamentals. Sterne has constructed a stirring narrative, whilst tying technology, techniques and tools back to their impact on the business. The shape of the book, like an old-fashioned sales funnel, gives us confidence that what has been included is useful and important to business outcomes. He has charted some difficult territory – data quality, sentiment analysis as well as old and the new of social media.

However, there are a few gaps which could have done with plugging. The book title includes the term ‘optimisation’, and for my money, there were few examples of this. I would like to have seen a few case studies where businesses had ‘measured’ their social media efforts, understood that they were sailing off course, and taken corrective action to ‘optimise’ and get better results. By concentrating only on the ‘measurement’ side of the equation, the book feels slightly lopsided.

Secondly, a good number of the examples were B2C and featured the global brands. I know there is some value in featuring case studies from household names with household products, but I fear this is somewhat short-sighted. Many of the people reading this book will be responsible for business-to-business marketing programs, and will want to be inspired by these stories, no matter whether the brand name is global or not. In fact, some small business owners will see social media as an opportunity to compete on a more level playing field with their more well-resourced competitors, and would be interested to hear more about businesses who have taken on the goliaths in their sector.

Finally, I would have liked to see a greater exploration of Owyang’s maturity model. The model presented lacked depth and does not bear comparison with Stephane Hamel’s Web Analytics Maturity Model. Allied with this shortcoming, would have been a more detailed consideration of how to resource social media marketing – what type of people make great community managers, where to find them, how to define their responsibilities, what organisational guidelines should exist for using social media – all of these would have been important questions to answer.

These gaps notwithstanding, Sterne has written an important book about how we will measure social media and evaluate the results of our efforts. It deserves to be on the bookshelf of every serious marketer.

Making web analytics a continuous process

Dan — Mon, 08 Mar 2010 12:54:10 +0000

Vlad and Gus, forensic web analysts from the Web Detective Agency, are sitting in a canal-side café. Vlad is downcast after a tough meeting with a client.

Vlad: That was not a good meeting. We’re not going to get any more work from them, I am certain. This is the last thing we need. And they are going to tell everyone we are a complete waste of money.

Gus: Wait a minute before you jump into the canal, at least tell me why you think it went so badly. From where I sat, it went OK – not brilliant, but not terminal. How do you know this client?

Vlad: Dimitri put them onto us. I phoned Anna, the Marketing Director and we talked about what they were doing with analytics – not much at the time – and how they wanted to use analytics strategically throughout the business.

I suggested a workshop to agree their business goals and the most relevant KPI’s, and we had this workshop about three months ago. We also agreed that we would coach their in-house analyst to make sure that they were implementing a ‘world class approach to analytics’, as Anna calls it.

Gus: So why do you think the meeting went so badly?

Vlad: Because after seeing some immediate improvements in their conversion rate after the workshop, it has started to plateau, and we are getting the blame for it. And probably, Anna’s right, they are not ‘world class’ and we’ve let them down, and you know Anna’s well-connected and she’ll be tweeting about us right now – saying what a poor job we’ve done.

Gus: So, you didn’t say the workshop was actually a ‘disaster’. Assuming that went well, something has clearly happened between the workshop and now, right ? Who came to the workshop?

Vlad: Lots of people were down to come but there were quite a few no-shows. I don’t think Anna had spelt out why they should come along and what role they were expected to play. Plus, by the end of the workshop there were just so many website KPIs that they were going to measure and optimize.

Gus: OK, I can see not having some key people at the workshop not helping matters right now, but why did you let them apply KPIs to absolutely everything ?

Vlad: Anna was insistent. She said that they had just invested in a new analytics tool that could produce these wonderful reports at the push of the button. She didn’t want to tell her boss, who had agreed to spend over 30,000 Euros on this system, that her team was only looking at less than 10% of these fancy reports. I kept saying that she was going to overwhelm her colleagues but she wasn’t listening.

Gus: The web analyst guy, Alex, it seemed Anna was pretty tough on him as well.

Vlad: I think he’s been put in a really difficult situation. He sits in the IT department, even though he is supposed to be working for Anna and her marketing team. He’s been told that he’s responsible for raising the conversion rate and yet because he’s got an IT background and sits with all the other IT people, he gets dragged into other projects.

Gus: It seemed to be that there was something else going on. When Alex presented the results on the last split test they’ve did on the product page I thought Anna’s head was going to explode

Vlad: Just like in ‘Scanners’.

Gus: What, what are you talking .. anyway. No, I meant that Anna’s was clearly not expecting a test to produce anything less than a fantastic result. When Alex showed her the figures from testing the original version of the product page against one we had recommended there was just silence.

Didn’t you explain that part of the testing process is that you sometimes find out what doesn’t work, as well as what does?

Vlad: I thought I did, during the workshop. But the way she spat out the word ‘failure’ I don’t think she has really taken it in.

Gus: But I thought there were some positives – Alex showed that some KPI’s had improved.

Vlad: That’s right, but because we had run out of the time during the workshop, partly having so many KPI’s to agree, we never got round to setting achievable targets for each of the measures. So, even though we have got some improvement, no-one is going to get a pat on the back to say ‘you’ve beaten your target’ because none were set.

Gus: Just thinking about the web analyst Alex, he spent most of the meeting going through all the reports that he was generating. He’d clearly no time to spend on investigating or playing the detective.

Vlad: Well, we’re supposed to be helping him with that area. But when I went to see him 6 weeks ago to help him delve a bit deeper into the data, all he was doing was getting the reports set up and scheduled to be delivered to just about everyone in the building.

Gus: And did you notice the reports?

Vlad: No, what do you mean?

Gus: What did you think about how the reports looked ?

Vlad: They looked alright, apart from there being far too many.

Gus: Vlad, they looked awful, all numbers in densely packed tables – not a graph to be seen. No series of traffic lights – you know, red, green and amber.

Vlad: He was probably too busy to worry about what they looked like. But I suppose you’re right – no-one is going to take much interest if they look like my old stats exam papers.

Gus: And what happened to bringing in results from the visitor surveys?

Vlad: Mmmm. I spoke to Alex about this. A fair number of survey responses were pretty negative about the product imagery on the site, and he was so worried that Anna was going to have a go at the art director, Johann, that he said he had ‘forgotten’ to download the data from SurveyMonkey.

Gus: But that type of insight is exactly what they need to understand what to improve.

Vlad: I know, I know. There’s no way we are going to get out of this with our reputations intact. The next email I get from Anna will be her terminating our retainer.

Gus: Hold on Vlad. Look I’m sure we could have done a better job in the workshop of trying to get Anna to understand that analytics is a process that sometimes gives results which aren’t what we would like, but I am sure the situation can be salvaged.

Let’s arrange another meeting to go through the issues we’ve just discussed. We’ll explain that the results from her analytics project could be *even* better if she understood that analytics is more than just a technical process or solely about conversion rates. She needs to stop ‘blaming’ members of her team and start seeing this as a journey – with its ups and downs.

Don’t forget, with our help, Anna did get a big boost in conversion in the early days – her boss will still be thinking that she’s doing a great job – but we know that if we don’t get this process right now then six months down the line we will have a much harder job getting it back on track.

Vlad: Plus, we need to move Alex physically to sit with marketing team, and prioritize his reports with the marketing team, at least to give him some time to be explore what the data is telling him.

Gus: And don’t be so hard on yourself. Just because we’ve run a workshop with Anna and her team doesn’t mean that we can all just relax and assume it’s going to work. This is about making a process successful, not just analytics.

Vlad: You’re right. I just expected people to do what I recommended, and then follow the tips we gave them. I didn’t expect that we might have to remind them of what we’d told them.

Gus: I think if we forget to consider the people involved, then of course the process will suffer.

Vlad: You’re right. I’ll give Anna a call on Monday to set up a meeting.

____________________________________________________________________________________________________________________________

How to select website KPIs

Dan — Mon, 15 Feb 2010 20:34:05 +0000

Vlad and Gus, forensic web analysts from the Web Detective Agency, are in the office working out how to create some KPIs that will match their clients’ business objectives.

Gus: Hey Vlad, we have been helping our clients work out what’s been going on on their website for ages now, but how many of them have actually got a clear set of website KPIs for their business?

Vlad: Most have, but some haven’t. But each business will have different objectives

Gus: Correct, but there are some common business models, like e-commerce, lead generation, advertising, and customer support, where the KPIs are likely to be similar.

Vlad: So for e-commerce sites, what would the KPIs you’d look at be?

Gus: The list would go something like this:-

% of new to returning visitors
Bounce rates on key pages
Referring websites
Search keywords and phrases
Conversion rates
Average order value
Sales per visitor

Vlad: OK, OK, but I don’t believe that each business, whatever model they fall into, would want to use exactly the same set of reports.

Gus: No, you’re right, and that is why if we are going to help these guys, we need to show what they should be measuring, specific to their business

Vlad: I think I may have seen a few models that could help us, like those that Petersen, Jackson, Sterne and Kaushik have come up with

Gus: What can these models show us?

Vlad: Well, before looking at that we need to understand what is involved in the creation of a KPI.

I remember Jackson talking about a “core score” – you know, a measure of the whole organisation’s efforts. He talked about the positive effect on the UK Prison Service when it moved from measuring itself on the number of escapees to judging its effectiveness in terms of number of repeat offenders.

That switch meant that the organisation considered itself as part of a system, rather its role within the system.

Gus: That’s a good example, but most of our clients are in the private sector where being judged on their revenues and share price is the *only* number they care about.

Vlad: Yes, of course, somewill see the success in those terms, but think about a goal like “the creation of happy customer”. Creating lots of happy customers creates great financial results for the organisation, but also if you unpack the term ‘happy customers’ you’ll see there are lots of hard measures a client would need to monitor and manage – levels of repeat business, customer satisfaction, loyalty and so on.

And that’s my point, you might have a strategic goal, but in order to succeed at that objective you need to have a good number of operational or intermediate goals as well.

Gus: And what type of operational goals could those be ?

Vlad: Well, according to Jackson and the others, it’s all a variation on this theme

How well are we doing at reaching our target audience (or visitor acquisition)?
How well are we engaging our audience (or converting them to the opportunity)?
How well are we doing at activating them, or turning them into a customer?
And how well are we doing at retaining them or nurturing them?

But the important thing is how each business defines this ‘core score’ as well as these operational goals.

In the past I have used a KPI workshop to generate these goals, and out of these come measures for the website.

Gus: Who do you get to come to this workshop?

Vlad: I think there are three important sets of people to have come along

People who run departments and manage budgets
Those that are responsible for what appears on the site – whether that is product, or service description, or editorial
And finally, the detectives, us analysts who can help them with evidence – you know analysis and reports and, and so on.

Gus: Lock in the room and let them thrash it out then ?

Vlad: Not at all. I have always played the facilitator – you know armed with a the marker pens and the flipcharts.

The first thing you do is explain the Reach, Engage, Activate, Nurture model (I prefer Jackson’s model over the others). I always draw four boxes on a flipchart

Second, you ask everyone to brainstorm the questions they need the answers to to allow them to do a better job.

Gus: For example?

Vlad:

Where are the visitors coming from ?
What do people search on the site ?
Which products convert the highest ?
Why do people buy our products?
What is most profitable way to acquire new customers ?

Questions will be a mix of Reach, Engage, Activate and Nurture

Gus: Then, I am guessing, you read out each question from your workshop people and put them into the correct box on the flipchart

Vlad: Absolutely, and then you jointly agree a measure how that goal will be measured. Here, have a look at this sketch I have done.

Website KPIs

So for example, looking at most profitable way to acquire new customers, you would generate a report on different forms of traffic segmented by ROI.

Gus: Aren’t you forgetting a couple of things

Vlad: Like what?.

Gus: Firstly, you need to what action will be taken as a result of this KPI. There’s no point getting people to produce reports if no action will be taken.

Vlad: Agreed, 100%. And your second poist.

Gus: You should have some form of benchmark.

Vlad: What type of benchmark?

Gus: Well, I know there’s a lot of benchmark data on the web, but at least you should benchmark your own performance. As well as benchmarking your performance against previous results you need to set a target for improving each metric over a period of time.

Vlad: Agreed. The last point is that you need to really embed this process, and ensure you’re doing this on a continuous basis

Gus: And what about how the reports look, isn’t that important?

Vlad: Too important for a brief conversation, let’s come back to that one another time

__________________________________________________________________________________________________________________________

Setting website goals – part 2

Dan — Mon, 23 Nov 2009 11:22:54 +0000

Vlad and Gus, forensic web analysts from the Web Detective Agency, are reviewing the performance of one of their clients’ email marketing campaigns.

Gus: Their results just get better and better. Open rates are up, clickthrough rates up and so is the number of orders generated from each email.

Vlad: They have spent a lot of effort on testing different subject lines, offers and creatives. It’s starting to really pay off for them.

Gus: And who suggested they do this?

Vlad: Alright, Gus.

Seriously, you have really helped these guys. Without your advice and also encouragement they would still be complaining about how they never seem to make any money from their emails.

Gus: Well, if you are giving each other pats on the back, you need one for increasing the numbers of emails that actually got delivered. The work you did on checking the emails wouldn’t be caught by spam filters and checking that their email server hadn’t been blacklisted with ISPs made a massive difference as well.

Vlad: I just assumed that they were doing this already. Just commonsense really.

Gus: What it means is that each email sent is now generating 15 cents instead of the 5 cents we started with. Given that, I am surprised how relaxed Luca is about recruiting new subscribers to their email newsletter.

Vlad: Each subscriber clearly has a value, and yet he;s not very aggressive about adding new email addresses.

Gus: How could we demonstrate to him that he needs to get serious about recruiting more subscribers?

Vlad: I think the first thing we have to do is set up a goal on in Google Analytics, and start recording how many new subscribers they are converting. Next, we need to put a value on that goal.

Gus: Like the value we put on a catalogue request goal?

Vlad: Exactly. Let’s see if I give you the raw numbers you can work out how much each email subscriber is worth to them.

Gus: Hit me.

Vlad: OK. We know that they get 12,000 unique visitors per month, their average order value is 45 euros, the conversion rate is 3% for those who have subscribed to an email newsletter, and 2% for those who haven’t, and they pick up roughly 450 new subscribers per month.

Gus: So, the revenue they generate from their email subscribers is.. 12,000 times 3%, so that is 360 orders per month, multiply that by AOV of 45 euros you get .. 16,200 euros per month. Right so far?

Vlad: So far.

Gus: If we compare that with the revenue they get from the non-subscribers, 12,000 times 2% multiplied by 45 equals.. 10,800 euros per month. So the extra revenue they get from email subscribers is 16,200 less 10,800 equals 5400 euros per month. So if I divide the extra revenue, the 5,400 euros, by the number of new subscribers each month .. I have forgotten how many email subscribers they get each month?

Vlad: 350

Gus: 5,400 divide by 350 is 15.42 euros.

Vlad: So each email subscriber generates an extra 15 euros per month, or 185 over the course of the year.

Gus: That really makes it very clear how much each subscriber is worth to the business. I think when we show this, Luca he’s going to have to get a lot more focussed about getting more names ontothe email database.

Vlad: We need to talk through all the different ways he can promote the email newsletter, as well as looking at how we can entice more people to sign up.

Gus: I think creatively there is a lot that they could do to make the call to action more compelling, also having the email signup box at the bottom of the page is doing him no favours.

Perhaps I could get one of the designers I know to mock something up.

Vlad: Good idea. We could split test the new creative against the current one to ensure that this is the reason email signups are going upwards.

Gus: I’ll give the designer I know a call after lunch and brief her.

Vlad: Excellent.

____________________________________________________________________________________________

Related Posts

Setting website goals – part 1

Dan — Fri, 13 Nov 2009 16:19:23 +0000

Vlad and Gus, forensic web analysts from the Web Detective Agency, are just getting on their bikes to begin their journeys home after work. They are discussing one client’s obsession with his conversion rate – from visitors into purchasers.

Gus: I just don’t get it, all Johann talks about is his conversion rate. What is it this month? What was it last month? What are the underlying trends?

Vlad: It’s what drives his business. The higher the percentage of people that convert on his website, the richer he becomes.

Gus: I know that, Vlad, but bear in mind he takes a good proportion of orders over the telephone and via the catalogue. Butt because he can easily measure the conversion rate this is what he focuses on.

I know there are other things his call centre could do to increase average order values – a few competitions, special offers, phoning up best customers.

Vlad: I think for Johann, he been surprised by the growth in online orders and he’s started to pay less attention to other aspects of his business. What he needs to do is look at some other goals that he wants his visitors to accomplish.

Gus: Like what?

Vlad: Well, given the business started as a catalogue business, he should be focus on the number of people who request a catalogue via the website. Some people, given his target market, are never going to order online, but they may request a catalogue.

Gus: So, at least we can see how well the website is doing at generating a catalogue request, and which traffic sources are good or not so good for acheiving this goal. Ideally, we need that goal to have a value, so we can quantify these traffic sources and the keywords that generate catalogue requests. But is there a way of doing this ?

Vlad: I think even you could cope with the maths of this.

You take average order value, ideally of catalogue requesters, and then multiply it by the percentage of ‘catalogue requesters’ who convert into a customer.

Gus: My maths is getting better every day, I can now do long division!

OK, to prove my point, if Johann’s AOV is 90 euros and 10% of ‘catalogue requesters end up ordering, then the value of a catalogue request is 9 euros. Yes?

Vlad: I am indeed a great teacher. Yes, that’s it.

Gus: Good. Well tomorrow, I’ll set up a goal on his Google Analytics account and then get Johann to give me the AOV and response rate figures. It also means that he could set up a PPC campaign just targetting people who want to request a catalogue. This way he’ll have a target cost-per-request and good measurement on how well this campaign is performing.

Vlad: If we can get him to understand that there are other goals on the site for him to focus on then perhaps we can get him to see a bit beyond his damned conversion rate.

Gus: Agreed.

OK. Well I better get home. I need to change before going out with the girls.

Vlad: I like the sound of that. Do you want to come along?

Gus: Another time, Vlad. Another time. See you tomorrow.

Vlad: Yep, see you tomorrow.

The benefits of web analytics

Dan — Tue, 03 Nov 2009 17:04:18 +0000

Vlad and Gus, forensic web analysts from the Web Detective Agency, are discussing a presentation Vlad is going to give to a prospective client the day after tomorrow.

Gus: What time is the presentation?

Vlad: 10.30am at their offices on Stadhouderskade, near the Marriott.

I am not looking forward to it. When I set up the meeting on the phone with Niel, their Website Manager, he didn’t sound that committed to investing in web analytics.

Gus: Who else is attending ?

Vlad: He’s bringing along his Marketing Director, which means something I suppose.

Gus: Look, if you’re not positive about the meeting, then how do you expect to convince prospective client.

Vlad: I am positive, I’m just worried that this guy, Niel, hasn’t bought into it. Plus, I don’t think the first part of my presentation is very strong.

Gus: Well, let’s have a look.

[Vlad opens the Powerpoint presentation he’s going to use, and Gus starts flicking through the slides]

I think the problem is that you have one slide on what analytics, just a pretty dry definition from some textbook, and then you dive into the detail of what KPIs to measure.

You haven’t explained the benefits of web analytics; why their business needs it.

Vlad: What do you suggest I put in here?

Gus: Given that the Marketing Director is there, I would start with how web analytics will show them which traffic sources and keywords are profitable, and which ones aren’t.

Marketing Directors love working out where they can save money, and concentrate resources

Vlad: OK, so here’s my first one, if you don’t have web analytics you don’t know which sources of traffic and content is generating the most revenue, or profit.

Gus: Good, but phrased differently, what sources of traffic or types of visitors could be costing you money on your PPC campaigns.

Vlad: OK, so ‘save money and make more money’.

Another point is that your website needs to do accomplish lots of tasks – reaching out to new visitors, engaging them, converting them into customers and then retaining; and without analytics you are flying blind.

Gus: That’s right. These are the central functions of any commercial site, and if you aren’t measuring then you can’t be managing them.

Also you need to mention the impact on your brand of a poor online experience.

Vlad: What do you mean ?

Gus: I mean, online, your brand *is* your visitor experience, and if that sucks so does your brand. So for those businesses that have invested huge sums in promoting their brand, all that hard work can quickly be undermined if their website is unusable and frustrating. Like, covered in Flash, but ultimately doing nothing for the user, ‘flashturbation’ as Eisenberg calls it.

Vlad: That Bryan Eisenberg, he’s a funny man.

Also you could say.that without web analytics you can’t measure the contribution or impact that other departments within the organisation are making.

For example, you could evaluate the efforts of the merchandising department in encouraging visitors to view their products and add them to basket, across a range of categories.

Many merchandising departments won’t be used to being judged in this way, and it could lead to some friction, but with analytics you can make the relationship between their work and the activity on the website very transparent.

Gus: Yeah, I like that one a lot – it means holding the whole organisation to account for the website performance, not just marketing.

You could also judge the finance department in terms of speed of refunds against the number of emails receiving chasing refund payments, or in fact the warehouse on speed of despatch using surveys.

Vlad: Exactly, and added to that last point, if you don’t have web analytics you will be missing an important opportunity to understand more about visitors and customers. There’s not many other environments where you can find out as much about the way your prospects and customers behave and experience your service.

Vlad: Right, and we know that if you, the client, isn’t using web analytics, then you can be sure that their competitors will be.

Gus: Is that list of points enough?

Vlad: Yep, five is fine. Let me just read it back

Websites need to Reach, Engage, Acquire and Nurture

Gus: Yes, just explain that multi-purpose web site means to attract, engage, convert and retain

Vlad: Next one is that websites consume marketing expenditure and you have to know how effective this spend is.

Gus: So without analytics, you have got marketing expenditure, but no control

Vlad: Next one, ‘Is your website hurting your brand?’

The one after that is ‘Which departments are helping website performance?

And finally, ‘How much do you know about your online visitors?

Gus: That’s all of them.

Vlad: That’s good, I think there is a much better connection between what web analytics is and what KPIs you need to measure.

Gus: Let’s see if Niel and his Marketing Director think so too.

____________________________________________________________________________________________

Related Posts

What is Web analytics?

What is web analytics?

Dan — Tue, 29 Sep 2009 20:38:57 +0000

Vlad and Gus, forensic web analysts from the Web Detective Agency, have just arrived in the office and Gus is making them both a cup of coffee.

Gus: I left without saying goodbye last night, but you were in deep conversation on the phone.

Vlad: I know. I got an enquiry from the website, so I thought rather than wait around for them to reply to my email, I would give them a call.

Gus: How did you get on ?

Vlad: Well, it turned out to be an enquiry that we couldn’t help her with.

She was looking for a new search marketing agency to replace the ones they have currently got.

Gus: So why did she think we could help her?

Vlad: It happens a lot. Many people when we they read on the website, or when I tell them that we can help them increase their sales, they assume this means we’ll deliver more traffic to their site through SEO or PPC.

Gus: But don’t they understand what web analytics means ?

Vlad: Sometimes not.

Although it’s interesting sometimes, because once I start talking to these people who have confused us with a search agency, I get to ask them some awkward questions – like what’s your best converting keyword, or which page on your site has the highest bounce rate ?

It’s often at that point they understand that it’s what analytics is about – understanding what visitors are doing on your site.

Gus:I know you’re king of the clickstream, but don’t forget that there are other ways of understanding visitors.

For me, as well as their behaviour – the what – it is also about the understanding the visitor’s experience and to do this you have to watch them or to listen to them.

Vlad: But that not statistically significant data.

Gus: Sometimes that’s important, sometimes not. But when clickstream has identified ‘what’ the problem, what else is going to tellyou ‘why’ there’s a problem?

The best people to tell you more about the problem, and give you ideas of how to fix them, are visitors themselves.

Vlad: You have to remember that’s there a difference between what people say and what people do. Clickstream data gives you that truth, without the visitors having to vocalize their thoughts.

Gus: That’s true to a degree. But I think what we are trying to do is to identify variables that if we change, result in better website performance.

Vlad: What do you mean by performance?

Gus: It’s something you’re always talking about, that most websites don’t exist for fun and entertainment. They exist to perform a function; therefore performance is whether or not the website allows a visitor to fulfil the purpose that came to the website for.

Vlad: We agree on that. And what I would say is that once we have identified what needs fixing – however we do that – we can use techniques like splits test to confirm that what we changed has resulted in a measurable improvement.

Gus: Split what ?

Vlad: I’ve told you about these before. You set up one or more improved pages that are different from the existing page – and then you compare which page resulted in the most conversions, such as transactions, or email newsletter sign-ups.

Gus: OK. I see. And the benefit of that is…?

Vlad: It either proves or disproves your hypothesis as to whether you have improved on the problem page, or whether you need to go back to square one.

No more arguments, just hard numbers.

Gus: The bottom line is that web analytics is about improving the bottom line for a business.

Vlad: 90% of the time, that is exactly what it is about.