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!”

• Being clear about your goals
• Defining the KPIs and key metrics
• Defining the business processes
• Writing a challenging business requirements document

But selection is only half the battle to successful adoption, there is also the issue of actually implementing and getting business value from it. In the past I have seen systems implemented that work perfectly well but are not felt to deliver much in the way of value, simply because they are being underutilised (for whatever reason) relative to their potential. So, what are some tips for the successful implementation of analytic systems?

Write a plan
This may seem like a case of stating the obvious but the point that I would make is to write a type of business plan rather than a technical systems implementation plan. The plan needs to consider all the environmental and organisational issues as well as the technical issues.

On one level there are all the technical issues that need to be thought through such as how is the site going to be tagged or how are the log files going to be harvested? The devil is in the detail with these sorts of issues and careful consideration needs to be given to them upfront to avoid pain later.

One another level the success of the system implementation will also be judged on whether people are using it and whether they are getting any value from it. This is where the organisational and environmental factors come into play. Different types of organisations and functions are likely to have different attitudes to the adoption and use of analytical reporting systems. So what’s the plan to get them engaged on the level they need to be engaged at?

Find advocates

As part of the plan you may want to think about who your advocates are likely to be. These are the people who “get it” and will be willing adopters of the new system and the opportunities that it will offer them.

One of the mistakes I made in the past when trying to roll out a new analytical system was to try and roll it out on too wide a front. Too many people were involved and at the time it was like trying to push water uphill. Since then I have found that a useful approach is to find group of individuals or a function that are happy to be early adopters and work closely with them in the early phases to get the system actually being used in the business.

In the case of a web analytics’ system you may work with the advocates to get some of the more sophisticated features of the system up and running such as campaign tracking capabilities. Working with these advocates can then help you do develop useful case studies that you can use with the rest of the business in seminars, training courses and the like.

Deliver appropriate training

In order to get the most out of a new system everyone needs some training even if they are not going to be actually using the system itself. Training can often focus on how to “drive” the system itself, a bit like learning how to drive a car, but it⿿s also useful to coach people in what to do as a result as well, and that’s where the case studies you’ve built up with your early adopting advocates come in handy.

In many cases systems can “push” reports out to end users via email or on an intranet, which is great. But do they end users always know what these reports are telling them and what they should be looking for? I’ve argued in the past that an analyst adds value by adding interpretation but if this isn’t possible all the time at least think about how to coach end users in how to interpret the data that’s being delivered to them.

Demonstrate value and celebrate success

The proof of the pudding is in the eating and as I mentioned last time you only really know what you’re going to get with a new analytics system once you’ve got it in and up and running. It’s important that the value of the investment is demonstrated to the business. This may take different forms depending on the business but there should be some early winds that you can point to in order to demonstrate the success of the implementation. These may be new approaches to campaigning that are being implemented or improvements in conversion in certain areas. Whatever they are, add them to your library of case studies and use them to show the business value every opportunity you get!

Next week I’m heading over to Search Engine Strategies in New York to take part in the ClickZ track. I’m really looking forward to catching up with some old colleagues and hopefully meeting new ones. I’m also looking forward to taking a look at the search analytics scene and will reporting back on what I find next time round. Till then…

I’ve worked with a number of companies over the years helping them to select and implement analytical software including web analytics packages. Indeed I went through the painful process myself, implementing a tag based ASP solution across a dozen sites across Europe and integrating it in with other information and operational systems. As a result of the work I do with clients, I have heard lots of stories about why they want to move from their existing package to another one.

Some of the common themes that I come across are:

The system doesn’t deliver any value to the business…
The system sits in isolation and it’s difficult to get it to integrate with other data…
The system doesn’t appear to work properly and deliver what’s needed…
Nobody really knows how to use it…

These themes are usually the symptoms of a poorly thought through selection and implementation process for the existing system and may not always be a problem with the system itself. Often these issues stem from a lack of clarity about what is really needed in the business and also an underestimation of the associated effort required.

There can be a tendency in the process of selecting analytical systems to be led by the technology and the feature set. As I mentioned in the last article, funky overlays and pretty dashboards might make the data look good, but you also need to ensure that you can get at the right sort of data you need in the right sort of way.

Here are some thoughts about how to minimize the risk of being one of the “38%”.

• Be clear about your goals
• Be very clear about what you are trying to do online and the reason for the investment in the web channel. From this all else flows.
• Define your KPIs and main tracking metrics

Once the goals are visible, it’s possible to identify what the channel Key Performance Indicators (KPIs) are. In addition, there will also be a number of other important tracking metrics as well, which are not as strategically important as KPIs. It’s important to recognize that KPIs are not always metrics that come out of a web analytics package. In fact, in my experience they rarely are.

Define the business processes

As well as defining the important metrics, it’s important to map out and define the important businesses processes. How are campaigns managed? How is the site development process managed? The tools should help the business processes and you shouldn’t need to change your business processes to fit the tool, unless it is clearly a much better way of doing it.

Write the business requirements document

This is the document that goes out to potential vendors. The point of the document is to clearly set out your needs and to invite the potential supplier to articulate how their system can meet them. In my view this document should enable the best potential suppliers to shine through, so don’t be too prescriptive in approach. The kind of huge document that requires huge amounts of detail input on every single aspect of the system does not help at this stage. You hate writing them, the vendors hate responding to them and then you hate reading them. And even then, there is no guarantee that you will be able to see the wood from the trees anyway.

Keep the requirements document concise, clear and open. Invite vendors to make an effort to differentiate themselves. A vendor once told me that the best brief they had had from a prospect was “This was what we are trying to do, tell me how you can help us do it better” and that was it.

Get it down to two or three

From the various responses that you get, you need to whittle it down to two or three potential vendors to come in and pitch for the business. Don’t invite them all. Look for quality not quantity in the response documents that you get back from them. First of all, have they answered the brief? Have they thought about what you asked them to think about? Or have they merely cranked the document out of the proposal machine? Do they demonstrate an understanding of the business and how they can help? Does it look like they want your business more than the next person?

Test it out if you can

In an ideal world you should test at least your preferred system, or preferably two systems side by side to see them in action. Analysis and reporting systems are “experience products”, you only really know what you are going to get when you’ve already got it. Having real hands on experience will be invaluable in helping you decide whether this is the one from you and you are going to be one of the “62%” of organizations.

Selection though is only half the battle. The other half is getting the system in, working properly and being used to its potential. Next time, I’ll being sharing some thoughts on successful implementations. Till then…

My fellow columnist Jason Burby has written extensively about setting KPIs for different types of web properties and it’s worth reviewing those articles. I just wanted to add some thoughts and my perspective on the subject of developing KPIs and the dissemination of e-business insight around the organisation.

In an ideal world, all organisations will have clear goals and objectives. Flowing out from those goals and objectives will be the organisation’s Key Performance Indicator’s (KPIs). These KPIs are the measures by which the organisation knows whether it is being successful or not. In a similar way the e-business channel would have its own goals and objectives which would be aligned to the overall corporate goals. From the e-business’s goals and objectives would come it own KPIs which would have some resonance with the corporate KPIs.

For example, take a multi-channel retail banking operation with branches, a call centre and an online banking presence. One of its corporate objectives might be to improve the efficiency and cost-effectiveness of its customer service operations. This might have a corporate KPI attached to it like the cost per customer transaction.

In terms on the online channel, this corporate objective might manifest itself into a desire to increase the use of online self-service facilities. A KPI for this might be the number of transactions carried out on the online bank. By measuring the volume of transactions being carried out on the online bank against its targets, the bank will know whether it’s being successful not in improving the adoption of self-service facilities and reducing its cost per transaction.

One of the potential traps when developing KPIs is to have too many of them. It’s important to remember that KPI stands for Key Performance Indicator and so the focus of measurement and attention is going to be lost in there too many metrics being tracked. My view is that there should only be a handful of KPIs at a time, ie about 5 or 6. A small number of KPIs are easier to remember and to keep at front of mind. Too many KPIs become unwieldy and result in a lack of focus on the main objectives. If you have 30 different measures each with the same perceived priority, then there is a danger of “paralysis by analysis” as you wade through a sea of metrics trying to understand what is happening to the business.

Selecting KPIs is a strategic process but that doesn’t mean that they then become set in stone. They are dynamic metrics that can change from year to year as the goals and objectives of the business change.

I believe that there is a tendency amongst e-businesses to focus on metrics that come out of web analytics packages and call them KPIs. I heard an interesting comment at a conference earlier this year from a guy who used to run Amazon in Europe. He said “Be careful of what you count. What you count is what you get”. I think that this is a powerful piece of insight as it links the way that organisations are measured into the ways that organisations (and the people within organisations) will then behave. Are you more interested in increasing page views or are you actually more interested in improving customer satisfaction?

Another tendency is often to focus on KPIs that are outcomes rather than the causes of those outcomes. Shouldn’t the focus be as much on the drivers, as on the results? For example, on a commerce site a KPI is often “Sales”. The sales figures will also typically be part standard financial reporting pack. However, sales are an outcome or a result, what is it that is driving sales and should some of those drivers be KPIs?

So, what makes a good KPI? One succinct answer from Craig Schiff, a business performance management consultant, is that the qualities of good KPIs are that “they are strategic, they are actionable and they will make a difference in improving the business”. How do your KPIs stack up against these criteria?

Once KPI’s are agreed, understood and measurable, then I believe the process of data dissemination becomes a lot easier. The organisation is focussed around the key issues and knows what data they should be looking at or looking for. There are still obvious challenges in finding ways and means of bringing web data to life and that’s something I’ll be taking a look at next time round.