Analytics Strategy

This article, written by Neil Mason, was originally published on Clickz.com and is republished here with permission.

As I write this I’m on my way back to the UK from the Emetrics Marketing Optimisation summit in San Francisco. After three days and having watched about 15 presentations this is probably the first chance I have had to reflect on what I’ve seen and what I have learned. The first thing that strikes me was the breadth of content that was covered. I went to presentations ranging from the “emetrics ecosystem” to usability and from testing and experimentation to social media measurement. The second thing that stuck me was the quality of the material and the presenters. In a show like this it’s possible to hit a duff presentation or two but looking back at my notes, all the sessions I attended were spot on.

So what did I take out of the conference? One theme that came through was there was a lot more evidence of organisations using integrated measurement strategies than I have seen before. More organisations were showing how they use a wide array of tools and techniques to understand the effectiveness of their digital marketing programmes. Voice of the Customer methodologies such as onsite feedback and surveys are the norm, most people are using testing and experimentation approaches and the use data mining and predictive analytical techniques is on the increase. Text mining tools are being used on verbatim comments from onsite surveys to extract the core essence of what is being said.

The stage was set on the first day with a keynote presentation “Competing with Analytics” from author Tom Davenport. There were some interesting things that Davenport said that set the tone for the conference. “The planets are aligned for analytics” he said, meaning that all the necessary components for organisations to adopt and deploy analytical capabilities are being put in place: Data, Enterprise, Leadership, Targets and Analysts. “Using analysis is good, competing on analysis is better” summed up the need to be able to move from insight to action. There is no point knowing stuff if you don’t do anything about it. He described the five stages of an organisations analytics capability from being “Analytical impaired” at the low end of the scale to being “Analytical competitors” at the other end. Organisations such as Harrahs and Marriott ion the US and Tesco in the UK use analytics as a source of competitive advantage.

Another stand out presentation was from Tim Goudie from The Coca Cola Company. Tim described Coke’s journey from the early implementation of their web analytics platform through to the development of their whole measurement framework. Goudie told us that “Metrics are ridiculously political; there is no such thing as a neutral metric”. Once you begin to measure things, then your are likely to start to change behaviour.

Other sessions I attended confirmed my belief that measuring and understanding the impact of social media is still in its infancy. Metrics and measurement frameworks are still in development, debates still stage about the meanings of terms like “engagement” and so on. Fellow columnist Jason Burby reminded us that when it comes to social media measurement of the importance of defining what success looks like in terms of key behaviours and that whilst the activities may be different the underlying measurement processes are the same.

A great take out from the presentation by Ebay was the use of “home visits” to better understand the user experience. I have seen this technique used by consumer packaged goods companies where people from the company visit consumers in their own homes to see them, using their products in real life. This was the first time that I had heard of this approached being used by an internet company. Executives from Ebay would visit users in their homes to understand the context within which the site is being used, revealing more insight into what really goes on than a standard usability test would.

Jacob Nielsen though showed what you can get out of usability testing in the laboratory. He asserts from the many tests they’ve completed over the years that task completion rates are going up and stressed the value of using testing early on in the development process. Oh and by the way “most people tend to ignore junk on websites”.

So this industry is now more than just what comes out of a web analytics tool. It’s about having a range of tools and technologies embedded within strong business processes. As Avinash Kaushik told us it’s about “multiplicity, flexibility and agility”. The planets are indeed aligned for analytics.

This post originally appeared in Applied Insights’ events section. Foviance acquired Applied Insights in November 2008, with Neil Mason joining us as Director of Analytical Consulting. As part of this acquisition, we’ve incorporated Applied Insights’ events list into our own.

At this year’s annual Web Analytics Congress in Holland, Neil delivered a keynote presentation on Marketing Optimisation and Predictive Analytics.

This post originally appeared in Applied Insights’ events section. Foviance acquired Applied Insights in November 2008, with Neil Mason joining us as Director of Analytical Consulting. As part of this acquisition, we’ve incorporated Applied Insights’ events list into our own.

This year Neil was invited by Jim Sterne to be the conference chairman and a keynote speaker the Emetrics Marketing Optimisation Summit in London.

As well as fronting up the proceedings over the two days and trying to keep the conference (and its speakers) on track, Neil delivered a keynote presentation looking at the development of web analytics and marketing optimisation practices within organisations called: “To Marketing Optimisation and Beyond!”

This article, written by Neil Mason, was originally published on Clickz.com and is republished here with permission.

The classic digital marketing processes are acquisition, conversion and retention and so far in this series on digital marketing optimisation I have been looking at the components of successful optimisation strategies when it comes to acquiring traffic on the site and then converting it. In this final part of the series, it’s time to look at optimising retention marketing activity. My own definition of retention marketing is:

“Retention marketing is the art and science of converting someone twice without the pain and cost of acquiring them twice”

It’s not just about getting someone to buy again (or whatever the conversion action is) but using what you know about them to improve the chances of converting them again without having to go through the whole acquisition process again.

But somehow “retention optimisation” doesn’t quite sound right and I prefer to think of it in different terms. What we are trying to do once we have acquired a customer is to optimise their lifetime value and so I tend to think about this process as “customer optimisation”. How can I optimise the return on the investment that I have already made in acquiring that customer in the first place? What data, tool, technologies and processes do I need?

Classically we tend to think of email when it comes to retention marketing channels and used well it can be a powerful retention tool. However once you have transacted with a customer there are multiple touch points that can be used to increase the chances of them doing business with you again; the call centre, the store, the site etc and what is required is a view of the customer that straddles these multiple channels. This much easier said that done, especially for organisations with legacy systems that have been developed over the years. Often data on customers can sit in a number of disparate systems and it can take a huge data cleaning and integration effort to get the data into shape and fit for purpose.

So having good quality data is important but what you then do with it is pretty important as well. As I said earlier, what we are trying to do here is to increase the likelihood that the customer will transact with us again without the cost of repeated acquisition. What we want to improve is the “expected customer lifetime value”. The way to do that is to be in the right place at the right time by being relevant and timely.

Being relevant is about sending out the right kind of messages, whether it is in an email or on the site. Segmentation is a way of increasing relevance. Personalisation is a way of increasing relevance. These techniques, which may be manual or automated, are leveraging the insight that you have about someone to present them with more appropriate and relevant. These techniques do not necessarily have to be sophisticated to be effective, and in the early days of your customer optimisation programme being over elaborate can undermine the process. Remember, a key component of any optimisation programme is to have the ability to execute. That means that you need to ensure you have the processes and tools in place which allow you to act, measure and react.

For example, if you decide that you want to improve relevancy by having a segmented email marketing programme instead of having a single email that goes out every month then you’re probably on the right track. However, if your resources and processes are geared up around just sending out on version of one email every month, then it’s going to be a major step up to implement a segmented email marketing programme whereby different groups of customers will get different versions of different emails at possibly different times. You will need to have a more sophisticated email systems that can handle segmented email marketing programmes, your database will need to be more extensive and robust, you will need to invest in more copy and creative material and your processes will need to be more rigorous.

So it will pay to walk before you start to run and to look for the low hanging fruit. In terms of customer optimisation, I think that the most critical point is getting someone to make the second transaction. Generally there is a “friction curve” that needs to be managed. The steepest part of the curve is in the early days of your customer relationship. The more times that someone has transacted with you the more likely they are to do it again in the future; the friction isn’t as high. Getting them to repeat for the first time is the hardest part. So this is a special case in the programme and is when timing can be vitally important. The time when people are most likely to transact with you is just after they last transacted with you, so for this special group all about “recency”. The time to get them thinking about the next transaction is just after the first.

Ultimately profitable businesses are built on profitable customers and repeat customers tend to deliver the majority of the profit. Acquisition and conversion optimisation are essentially components of a success marketing programme but optimising long term customer value is key.

This article, written by Neil Mason, was originally published on Clickz.com and is republished here with permission.

Last time in this series on digital marketing optimisation I looked at the process of campaign optimisation. But as the saying goes: “You can lead the horse to water, but you can’t make it drink”. Whilst your efforts make people aware of your brand and to give them a compelling reason to visit might be fully tuned, it’s what happens when they get to the site that makes the difference between online success and failure. So today, let’s take a look at conversion optimisation.

To start of let’s be clear about what we mean by conversion optimisation. To be more precise what do we mean by conversion? To optimise effectively you need to optimise some defined outcome and it’s important to clear about what the desired outcome is. For example, if you have a site where people can either complete the transaction online or pick up the phone and call the call centre, which one of those outcomes do you want to optimise? It may not be possible to optimise both. If you optimise the site to increase the online bookings you may find that the process doesn’t work for those people who prefer to research online but transact over the phone. If you optimise the site to generate call centre volume, it may mean that you have to incur extra costs in the call centre to deal with the extra demand. So, clarity of purpose is an important ingredient in the conversion optimisation process.

The conversion process is not one process; it’s a series of mini-conversion processes. Each of these processes needs to be optimised. You need to chunk the problem into bits. What is right for you will depend on your site and what you are trying to achieve. Fairly generically, the main micro-processes are: Land, Browse/Search, Engage/Transact. Part of the analytical approach to conversion optimisation is to identify where it hurts most. Which part of the conversion process seems to be causing the most problems? However, this also has to be balanced against the “ability to execute”. Which parts of the website can you change or impact in which timescales? It may be that the checkout process is identified as requiring optimisation and tuning but it may be that it will take 3 months before you can get the development resource to make the changes. In the meantime you be able to change some other parts of the site more tactically.

In an ideal world it would be great it is was possible to make changes to all parts of the site with relative ease and to measure the changes to get to the optimal result. There are technologies and platforms out there that are making this easier to do but in many cases the reality is that real choices need to be made in terms of impact versus time to effect. In these cases my instinct is usually to start at the beginning and look at landing page optimisation. Landing page optimisation is something that increasingly is becoming easier to do and it’s an area where there is always the potential of a high impact.

For me one of the biggest challenges in digital marketing is getting someone to do something twice. A lot of people who visit a website only visit it once, they only look at one page, and they stay for less than a minute. If they buy or transact, they only do it once. So how do you get them to do it twice? Look at that second page? Stay for that second minute? Make that second visit? A lot of subsequent behaviour is determined by the user experience on the first page of the first visit. The first page of the first visit needs to generate the momentum that ultimately leads to a successful outcome.

The growth of testing and experiment systems such as those provided by brands such as Optimost and Offermatica have made page optimisation processes a lot easier. They enable some of the hurdles associated with enacting change to be overcome and provide a systematic way to understand how to improve con version. They are not the only tool in the toolbox and conversion optimisation is an area where a wide array of data and services can be used in a holistic way to understand and optimise the user experience. Good site analytics gives an insight into the effectiveness of the micro-conversion processes, surveys help you understand the level of satisfaction with the user experience and usability testing tells you (warts and all) what processes works, which ones don’t and why.

Having gone through the process of acquiring prospects and converting them into customers, do you want to go through the pain and cost of doing it all again? I thought not. Next time we’ll take a look at the process of managing your investments in acquisition and conversion through optimising your retention marketing processes.

Till then…

This article, written by Neil Mason, was originally published on Clickz.com and is republished here with permission.

In this series of articles I am looking at the subject of digital marketing optimisation. At its core optimisation is a combination of philosophy, process, data and technology. You have to want to optimise and you have to be able to go through the “test, learn and adjust” cycles quickly enough to make a difference. Last time I outlined the approach to breaking down the digital marketing process into its main components of acquisition, conversion and retention and looking at the optimisation challenges within each of these areas. This week I look a deeper look at the first of these processes: acquisition.

Campaign optimisation is probably the most established optimisation practice. Acquisition marketing has been the corner stone of digital marketing for many years and a number of us can probably remember the mantra of the dot-com days when all the focus was on “eyeballs” and getting as many of them to the site as possible, often with little regards to what happened when they got there. Things have moved on a bit since then. Thankfully we now take a more sophisticated approach to measuring and optimising our acquisition marketing efforts. However, I think that there is still a way to go as campaign optimisation is a multi-level problem and once you have cracked one level then it’s time to move on to the next.

I think campaign optimisation operates on three levels:

• In-channel optimisation
• Cross-channel optimisation
• Multi-channel optimisation

In-channel optimisation

This is the process of optimisation within a single digital marketing channel (ie PPC search, display advertising etc) and is in effect what we have been doing for years. Over that time we have become more sophisticated and we have seen the development of solid processes and the deployment of analytics and technology to help automate the optimisation processes. Within PPC search for example, there have been bid management tools around for years and a whole industry discipline has grown up around the optimisation of PPC and organic search marketing activity.

We have also seen the technology evolve to the extent where campaigns can be optimised in almost real time through automated algorithmic approaches like of SearchRev in the case of PPC marketing or DART Adapt for display advertising. We’re almost at the point of “fire and forget” where once the initial deployment has been made and the system is calibrated, little manual intervention is really required.

That’s all great, as long as you are optimising against the tight thing and that you’re correctly attributing value to the right behaviours. A good case in point again is PPC search marketing. Classically PPC optimisation is done at the keyword level looking at the clicks that generated the conversions and adjusting bid-management strategies accordingly. However, we know that in a lot of cases it takes multiple touch points to generate a conversion, and so value needs to be attributed all along the acquisition value chain rather than just against what happened at the end.

Cross-channel optimisation

“In-channel” optimisation is relatively straightforward. It usually involves a single piece of technology and the data to drive the optimisation process is generally sitting in one place. The problem is that marketers rarely market using a single digital channel. They use a combination of different channels often at the same time to achieve their goals. Display advertising may be used for generating awareness; search marketing may be used for generating response and so on. Each of these channels might be optimised in their own right but what about the impacts of one digital channel on another?

A classic example of this was when one of my clients went through a process of optimising their search marketing and display marketing independently from each other. They came to the conclusion that their display advertising wasn’t working hard enough for them, whereas their search marketing was. They switched investment out of display marketing into more search marketing and then promptly saw their search volumes drop. Up until this point they hadn’t appreciated how much display advertising was driving search volumes because they were optimising within the channel and in the case of display they were optimising against the wrong thing. They hurriedly reverted to their previous approach and saw search volumes recover. But they were none the wiser. They knew there’s a relationship between search and display (and their other channels) but they didn’t know how to optimise it and that was because their data was all over the place.

The key to cross-channel optimisation is data integration. It’s necessary to have the different channels activity and responses in the same database. This might be achieved by using a single campaign management system, by deploying a universal tag or by integrating the data in a separate system. Either way, once the data is in the same database, it’s possible to look at the effect of different channels on each other and on conversions and then to be able to optimise accordingly.

Multi-channel optimisation

Increasingly digital or online marketing doesn’t sit in glorious isolation from an organisation’s other marketing efforts. It’s a multi-channel world and businesses are looking to integrate their online and offline marketing strategies and tactics. So the problem becomes broader. It’s not just about how do I optimise my search marketing, or how do I manage all my digital channels, it’s about how do I optimise the total marketing mix? Online and offline? Search, display, TC, radio, print – the lot? This is the strategic marketing optimisation problem that companies are beginning to wrestle with. Again one of the main challenges is data integration.

Online and offline marketing data generally have quite different characteristics which means that it’s not necessarily easy to integrate them. Online marketing data is usually at the cookie level, tracking an individual user (or more precisely, an individual device). Offline marketing data is usually at the “market” level ie starting at a country or a regional level and down to individual stations, publications and so on. User level data and market level data can be difficult to marry up and relate to each other. With the development of geographical profiling from IP addresses it is getting easier but the analysis techniques will then also be different and are more likely to resemble the inferential modelling techniques of offline marketing rather than the “direct cause and effect” techniques used in the online world.

So, campaign optimisation is a multi-tiered problem. It’s like going on a hike up a mountain. When you get to the top of one summit, another one comes into view and on you go. If you’ve just about got your search engine marketing cracked, it’s time to move on to the next problem!

Next time I’ll be taking a look at conversion optimisation. Till then…

This article, written by Neil Mason, was originally published on Clickz.com and is republished here with permission.

In my last column I looked at the concept of “data driven marketing” and outlined what I believe to be the for key components for successful data driven marketing, namely: Philosophy, Processes, Data and Technology.

At its heart data driven marketing underpins the drive towards optimisation but optimisation is one of those words that’s used a lot but often means different things to different people. A bit like segmentation, engagement and so on. So in this series of columns I thought I would offer my take on optimisation, what is means and how it can be achieved.

I first came across “optimisation” as part of my college degree course. It was presented as a mathematical problem; how do I maximise a desired outcome given a set of certain constraints? This is the way I have tended to think about optimisation ever since, particularly when it comes to marketing optimisation. My desired outcome will be things like orders, revenue or profit and my constraints will be things like money, time and resources. The challenge is to maximise the return on investment in marketing.

The problem with this purely mathematical approach to optimisation is that often the problem is too large or complex to solve easily. You need to understand and define the relationship between all the variables (such as sales and advertising) and then use the mathematics to determine the optimal allocation across the various inputs. The challenge is that there are often too many variables and the relationships are too complex to be solved easily. The approach then is to chunk the problem down and iterate towards a solution. This is where the components of data driven marketing come into play.

In the digital world the ultimate goal of many marketers is to maximise customer lifetime value and to allocate the resources available in such a way to achieve that. The issue is that things like customer lifetime value can be difficult to define and to measure and marketers may not have the ability to efficiently manage all the resources appropriately. The intent may be there but the ability to execute may not. As a result we break the overall process into smaller processes and we need to begin to think about how we can optimise individual separate processes rather than the complete value chain in one go.

We already think about digital marketing as three separate processes, namely:

• Acquisition
• Conversion
• Retention

Sometimes these processes can be too separated with little joined up thinking between the three. Having said that, for the purposes of optimisation and given the constraints of data and technology, it probably makes sense still to use them separately as the basis for our optimisation strategy. However, it is useful to bear in mind that what we are trying to so ultimately is to maximise the allocation of resources and investment across the whole customer lifecycle.

So what problems are we trying to solve? In acquisition we are trying to optimise our campaigns to increase the propensity of people to visit a website and engage. This is an area where there has been a lot of focus over the years and where technology has made a significant impact in either allowing marketers to iterate through the cycles more quickly or where the technology effectively automated the optimisation process. However, I think that the goalposts are moving and the problem set is changing and increasingly marketers need to be looking to optimise in a different way.

When it comes to conversion what we are trying to do is to increase the likelihood of some desired outcomes, whether that is an order, a registration and download or telephone call. Here there has been a lot more attention given over the past couple of years and where analytical technology has evolved to help marketers understand how to improve site architecture and design. There is more work to be done in this area though and the technologies and the processes to manage them need to be more widely adopted.

With retention marketing what we are looking to do is to increase customer value. Here we are talking about maximising on the investments that have already been made in acquisition and conversion so that we don’t have to make those investments again. With a few notable exceptions I don’t think that many organisations are focused on this area at the moment.

Over the next few columns I’ll take a look at optimisation is more detail, looking at acquisition, conversion and retention and examining the philosophy, processes required and data and techniques available to maximise the effectiveness these individual processes. Till then…

This article, written by Neil Mason, was originally published on Clickz.com and is republished here with permission.

I was recently asked to put together a workshop session on “Data driven marketing” for a class of digital marketing students. Part of what I asked to talk about to explain what data driven marketing is and how to go about it. In pulling the material together for the workshop I came to the conclusion that there are four key components for successful data driven marketing, some of which are obvious and some perhaps less so.

The four components are:

• Philosophy
• Processes
• Data
• Technology

Philosophy

This is the most important component I believe. To be successful at data driven marketing, an organisation needs to have the right culture and philosophy. At its hear t, data driven marketing is about continuous optimisation and iterative improvement. It’s the deployment of a “test, learn and adjust” philosophy. However, you can have the best data and technology in the world (see later) but if there is not the desire to act and to change, then the data and technology are only providing interest as opposed to insight. Organisations must have the “desire to act”.

At a Web Analytics Wednesday I attended in Berlin last week a lot of the talk in the networking session was not about metrics and systems but about how do you embed analytics within organisations? The biggest challenge often facing analysts is getting support for the development of their programmes because culturally the organisation doesn’t have a philosophy of measurement and accountability.

Processes

If “philosophy” is about the desire to act, then “processes” is about the ability to act. More specifically it’s about the ability to execute and then to react. These processes involve the management of the technologies and also the management of the decision making. Processes will include building “measurement” into the marketing development process for example, so that there is no question that new campaign won’t be tracked properly or that new content on the website won’t be tagged. It also involves ensuring that a feedback mechanism is in place that enables trends to be identified and changes to be made in the appropriate timescales.

One example I had in the past demonstrated to me where a potential desire to act was inhibited by an inability to execute. We did a piece of segmentation analysis for a retailer to feed into their email marketing programme. In the analysis we identified a number of distinct groups of customer with different purchasing behaviour that could be marketed to in a more customised way. We also identified some key timing mechanisms that could potentially double the customer’s propensity to buy again. Despite this insights the segments were never deployed operationally because the retailer didn’t have the resources and processes in place to develop and deliver more targeted email marketing programmes.

Data

Data is of course a vital ingredient in the mix, but it is the organisational culture and processes that provide the recipe for success. Good quality data is important and attention must be paid to getting the numbers right. People are reluctant to make decisions if they don’t have any faith in the data.

Also data driven marketing needs integrated data rather than data sitting in silos. Often within organisations different types of data sit in databases and different functions may have ownership of different data. For data driven marketing activities to be effective, the different data sources need to relate to each other. To understand and optimise marketing across channels, the data from different channels (PPC search, display ads, email etc) needs to be in the same place, whether that be in a web analytics systems, a campaign management system or both. In addition data needs to be managed across the life cycle of the customer, for example by ensuing that data on how customers are acquired can be analysed with the customers’ long term value or profitability.

Technology

Finally the technology is the enabling component. It is the technology that allows you to execute and react either over the duration of a planning cycle or even in real time. I don’t think that technology can make up for deficiencies in the philosophy and processes, though if you have the right apporach and procedures, you can make progress even if your technology is not the most effective. Good technology enables you to cycle through the processes faster to the point where real time optimisation is possible. Like the data, the technologies should be integrated and allow the loop to be closed between insight and action.

So, the core ingredient of data driven marketing is good quality, integrated data. The technologies are the tools but it is the combination of the organisational philosophy and strong processes that will provide the recipe for success.