Analysts largely agreed last year that web analytics solutions offered on an ASP basis were set to overtake the server-side (mostly log file analysis) versions which had previously dominated (Webtrends being the old stalwart most people know).
These ASP solutions, provided by people like Clickstream, Intellitracker, Redsherrif, WebtraffIQ, Webtrends NetIQ, Digimine, Manticore, Nedstat and Webside Story, use "browser-based" or "client-side" technologies ranging from simple remote calls to another server for an image through javascript to java code inserts. The data is recorded to the ASP's servers and access to reporting and analysis provided by a web interface to those servers.
2 things I'd have to say on the above:
1. Do people agree that ASP solutions will emerge as the way to do web measurement and analysis? I think they will as they are preferrable for a number of reasons....
2. Anyone used SpeedTrap? They're interesting for a number of reasons (video replay of users' visits among others) but, with reference to the above, they appear to be a kind of hybrid of 'browser-based measurement' - using a java applet installed on the user's machine - but NOT an ASP. They deliver a server to their customers which collects the data and provides the platform for reporting and analysis. But this server/infrastructure is then the customer's responsibility not them as an ASP.
Ashley,
Haven't come across SpeedTrap but I came across an impressive offering from UK-based Touch Clarity (http://www.touchclarity.com/) last year. It struck me that the ASP route was particularly useful if you anticipated switching platforms at some future point or if you had external groups who needed access to tailored reports (such as partners and suppliers).
I think you're right to identify ASP web analytics solutions as an important part of the industry, indeed, WebAbacus offers a hosted solution alongside our more traditional software licence sales. Customers benefit from reduced set-up costs and quicker access to business intelligence. But I would make the following points:
1. ASP solutions don't necessarily mean client-side technology solutions. I don't have a particular axe to grind here (we offer both server- and client-side analytics), but several of our hosted customers do use log files as the basis for their analytics. Debates about the accuracy or otherwise of log files notwithstanding, this does offer a very quick (arguably even quicker than a tagging-based solution) route to analytics. The client makes the logs available over an FTP or HTTP connection, we retrieve the logs automatically and process them. Set-up time can come down to a couple of horus, after which full historical data is available.
2. Hosted or outsourced solutions don't suit everybody. Whilst we have plenty of customers who are happy to have their web analytics hosted for them, we have plenty who could never go down this route. This latter group tend, in fact, to be our more sophisticated customers. They want to perform back-end data integration with their CMS or customer database to enhance their reporting; they also want complete control and ownership over the data (many of our customers consider web usage data to be company confidential material). Additionally, many of our customers are performing analysis of internal company networks - i.e. Intranets - where a 3rd-party-hosted solution would not be appropriate, again for security reasons.
As I said, we offer both licensed and hosted models to our customers, so I don't have any vested interest. I think the most important thing for us has been to be able to offer the customer a choice.
I agree that for websites ASP is probably the best route forward for most people as it offers a quick and easy solution.
BUT the big issue I find is that we all treat the website as a stand-alone enitiy rather than an integrated part of the business. The true value of a website is what it does for the overall business - and this is what needs to be measures. So your webabacus correspondant is right to my mind.... website results integrated with business results is what matters. We all need to go and learn about Cognos, Business objects and the other BI tools and how to integrated web results into them!
I agreee, an ASP will continue to make ground from the server-side solutions. The main driver is that people are realising that web data is complex, to do it well takes a lot of time and money. An ASP can deliver economies of scale (the development work is spread over their Client base) and reduced risk.
Regarding SpeedTrap my understanding is that it is primarily a usability tool. Good for understanding one visitor's experience at a time.
One of the problems with tracking is that a diverse number of tasks are covered by one name. I belive tracking will diverge into distinct areas of speciality, two of the most important being:
1. Those develivering database solutions (the abilty to improve the targeting of outbound communications based on form data, purchase and viewing history)
2. Those that focus on path analysis and usability (often using agregrate, unpersonalised data)
On 11:20:39 28 January 2003 Ashley wrote:
>Analysts largely agreed last year that web analytics
>solutions offered on an ASP basis were set to overtake the
>server-side (mostly log file analysis) versions which had
>previously dominated (Webtrends being the old stalwart
>most people know).
>
>For example:
>Have a look at this article = http://www.aspnews.com/news/- >article/0,2350,4191_1183081,00.html about this Aberdeen
>report - http://www.aberdeen.com/ab_company/hottopics/weba- >nalytics2002/default.htm (table of contents gives more
>detail)
>
>These ASP solutions, provided by people like Clickstream,
>Intellitracker, Redsherrif, WebtraffIQ, Webtrends NetIQ,
>Digimine, Manticore, Nedstat and Webside Story, use
>"browser-based" or "client-side"
>technologies ranging from simple remote calls to another
>server for an image through javascript to java code
>inserts. The data is recorded to the ASP's servers and
>access to reporting and analysis provided by a web
>interface to those servers.
>
>2 things I'd have to say on the above:
>
>1. Do people agree that ASP solutions will emerge as the
>way to do web measurement and analysis? I think they will
>as they are preferrable for a number of reasons....
>
>2. Anyone used SpeedTrap? They're interesting for a number
>of reasons (video replay of users' visits among others)
>but, with reference to the above, they appear to be a kind
>of hybrid of 'browser-based measurement' - using a java
>applet installed on the user's machine - but NOT an ASP.
>They deliver a server to their customers which collects
>the data and provides the platform for reporting and
>analysis. But this server/infrastructure is then the
>customer's responsibility not them as an ASP.
>
>Interested to hear any views.
>
>Ashley
Yes, I talked to Touch Clarity a year or two ago. What I remember as unique about their proposition was the ability to do real-time (sales) optimisation of web sites. They break the page up into various different ‘containers’ of content and then monitor which arrangement of content is most effective at driving sales. The system can then automatically optimise the page over time – one configuration might work better in office hours and another at the weekend for example.
It seemed to me this would work best for key pages within the site e.g. homepage, registration pages, checkout processes etc. I guess the concern from a site owner’s point of view is – am I prepared to outsource these mission-critical pages?
You make some very valid points. On point 1, it is true that log files are often the best place to start in order to quickly get some usable data, and, as you say, have access to historical data which is not possible with the ASP approach. I guess I’d still say they were a good place to start for a ‘quick and dirty’ piece of insight but quite possibly not the best longer term solution. A dual offering, as you have, makes absolute sense, particularly if you do want to migrate from one to the other.
Your point 2 is valid, though I have found recently that companies seem to be relaxing somewhat on their outsourcing fears. Transactional data and customer data are, understandably, very carefully monitored and guarded. Site activity data companies are less precious about now. With the advent of web services and the like I think this trend will continue – people will get more used to (and more comfortable with) outsourcing services where there is a good SLA in place and the firm is reputable, has a good track record, good clients etc.
I understand your point about the sophisticated guys wanting to keep control of the data in order to integrate with other sources - typically into a data warehouse for subsequent reporting and analysis etc. However, all the ASP solutions that I know of offer the raw data back to the client if they want it (as CSV, Excel etc.). I know of several instances where the more sophisticated organisations you mention use the ASP basically as a way to collect and ‘clean’ data for them. They see the ASP data as more accurate and more usable and more structured than log files (rightly so I think), so easier to integrate with their other systems. Although they buy the reporting and analysis functions with the service that’s not actually where the value lies for them.
It’s true that SpeedTrap seemed to develop a name for themselves as a usability tool. I suspect that is because of their mouse movement tracking reporting (uses a java applet client-side), their ‘how much of a form was filled in’ reporting as well as their feature to ‘replay’ a user’s visit. These features appeal to site designers and, indeed, anyone who finds it easier to look at pictures rather than grapple with data (which includes a lot of us, I suspect). Clicktracks (http://www.clicktracks.com) is also quite appealing and unique in the way it graphically shows where users have clicked on a page and you can browse the site like a real user and see the stats as an overlay.
That said, SpeedTrap does a lot more than this. It does all the usual stuff you’d expect. So it should do as the price tag is £10k+ putting it into competition with people like Site Intelligence and Clickstream whose proposition is more about more advanced and integrated reporting – going beyond basic site activity analysis into transactional and user profile data.
Your categorisation is interesting. I suspect you are right that there will be point solutions which essentially cater to different buyers – the database marketers want tools to do segmentation and outbound targeted marketing; the site developers and designers want tools to understand how they can optimise the site; the commercial guys want reporting to tell them about sales, conversion rates and so on. In some case a point solution will be sensible, in others a solution that can cover all elements.
It's certainly true that any self-respecting ASP will offer data download, though the data volumes can be very large, particularly when delivered as a CSV or Excel file. And whilst organisations are increasingly bringing clickstream data into their data warehouses, and using BI and analytical CRM tools to analyse this data, online behaviour analysis remains a specialised discipline, where specialist tools are needed. So really the reporting needs to remain in the analysis tool, and the extra data (such as customer records, or CMS data) needs to be imported in on demand. Apart from anything else, using an outsourced model for this would necessitate live database links over the Internet, which, whilst perfectly techincally feasible, is usually a bridge too far for most of our customers.
You are of course correct that what is most important is measuring how the web site, as *a* channel to market among others in most cases, is delivering value to the overall business. In most of the cases I've worked on the fundamental questions that (e-business) managers need to answer is 'what value are my web operations delivering? how can we confidently report and measure this? how do we optimise performance?
I agree that it would be helpful to understand 'traditional' BI tools which are of course more than capable of integrating data from web operations to perform analysis - I've done similar work with SPSS in the past. However, I'd add a few notes:
1. As Ian righly points out in his post at http://www.e-consultancy.com/forum/default.asp?v=1527&p=1 "online behaviour analysis remains a specialised discipline" so you still need the right skills and tools to measure the value of the digital channels. Having the data to analyse is one thing. Creating 'actionable intelligence' from it requires a deep understanding of the way the digital channels work - technically, for the customer, for the organisation etc. etc.
2. BI type analysis is one way to approach understanding value delivered. However, it misses out a whole load of other approaches which are perhaps 'softer' such as brand impact, awareness levels, customer experience and so on which need to be measured in other ways - focus groups, surveys, usability testing and so on. Again online can learn a lot here from offline but there are also specialist skills required to do the online stuff (be it online surveys, usability testing etc.)
3. Fortunately businesses are now beginning to accept that web sites deliver value to offline operations and cannot be judged solely on the direct (financial) performance of the channel in isolation. Web sites drive retail store footfall, web sites play an important part in the decision to buy a car etc. even if the transactions occur offline. One financial services provider I've worked with did a lot of business through the web site but 85% was transacted by phone following web research. They knew this simply by having a separate telephone number on the site. I think the whole area of measuring online-to-offline value creation is a very important one. It can also be difficult to do.
So I think there is nothing wrong with having a channel focus (digital channels) as there are specialist skills required. However, the really good people will appreciate how they fit into the wider business context.
The Web Analytics Beginner's Guide aims to be the first step in a journey that will see you become a fully-fledged expert in online analytics. A must-read for those interested in this area, free to Bronze members and higher.
The fifth annual Online Measurement and Strategy Report, carried out in association with web analytics consultancy Lynchpin, contains a comprehensive analysis of issues affecting the web analytics industry and valuable insights into the use of analytics and business intelligence tools.
CEO at Econsultancy
28 January 2003 11:20am
Analysts largely agreed last year that web analytics solutions offered on an ASP basis were set to overtake the server-side (mostly log file analysis) versions which had previously dominated (Webtrends being the old stalwart most people know).
For example:
Have a look at this article = http://www.aspnews.com/news/article/0,2350,4191_1183081,00.html about this Aberdeen report - http://www.aberdeen.com/ab_company/hottopics/webanalytics2002/default.htm (table of contents gives more detail)
These ASP solutions, provided by people like Clickstream, Intellitracker, Redsherrif, WebtraffIQ, Webtrends NetIQ, Digimine, Manticore, Nedstat and Webside Story, use "browser-based" or "client-side" technologies ranging from simple remote calls to another server for an image through javascript to java code inserts. The data is recorded to the ASP's servers and access to reporting and analysis provided by a web interface to those servers.
2 things I'd have to say on the above:
1. Do people agree that ASP solutions will emerge as the way to do web measurement and analysis? I think they will as they are preferrable for a number of reasons....
2. Anyone used SpeedTrap? They're interesting for a number of reasons (video replay of users' visits among others) but, with reference to the above, they appear to be a kind of hybrid of 'browser-based measurement' - using a java applet installed on the user's machine - but NOT an ASP. They deliver a server to their customers which collects the data and provides the platform for reporting and analysis. But this server/infrastructure is then the customer's responsibility not them as an ASP.
Interested to hear any views.
Ashley
Director at Contentment Ltd
03 February 2003 17:55pm
Ashley,
Haven't come across SpeedTrap but I came across an impressive offering from UK-based Touch Clarity (http://www.touchclarity.com/) last year. It struck me that the ASP route was particularly useful if you anticipated switching platforms at some future point or if you had external groups who needed access to tailored reports (such as partners and suppliers).
Andy
Marketing Director at WebAbacus
03 February 2003 18:09pm
Ashley,
I think you're right to identify ASP web analytics solutions as an important part of the industry, indeed, WebAbacus offers a hosted solution alongside our more traditional software licence sales. Customers benefit from reduced set-up costs and quicker access to business intelligence. But I would make the following points:
1. ASP solutions don't necessarily mean client-side technology solutions. I don't have a particular axe to grind here (we offer both server- and client-side analytics), but several of our hosted customers do use log files as the basis for their analytics. Debates about the accuracy or otherwise of log files notwithstanding, this does offer a very quick (arguably even quicker than a tagging-based solution) route to analytics. The client makes the logs available over an FTP or HTTP connection, we retrieve the logs automatically and process them. Set-up time can come down to a couple of horus, after which full historical data is available.
2. Hosted or outsourced solutions don't suit everybody. Whilst we have plenty of customers who are happy to have their web analytics hosted for them, we have plenty who could never go down this route. This latter group tend, in fact, to be our more sophisticated customers. They want to perform back-end data integration with their CMS or customer database to enhance their reporting; they also want complete control and ownership over the data (many of our customers consider web usage data to be company confidential material). Additionally, many of our customers are performing analysis of internal company networks - i.e. Intranets - where a 3rd-party-hosted solution would not be appropriate, again for security reasons.
As I said, we offer both licensed and hosted models to our customers, so I don't have any vested interest. I think the most important thing for us has been to be able to offer the customer a choice.
Ian
CEO at Logan Tod & Co.
04 February 2003 11:36am
I agree that for websites ASP is probably the best route forward for most people as it offers a quick and easy solution.
BUT the big issue I find is that we all treat the website as a stand-alone enitiy rather than an integrated part of the business. The true value of a website is what it does for the overall business - and this is what needs to be measures. So your webabacus correspondant is right to my mind.... website results integrated with business results is what matters. We all need to go and learn about Cognos, Business objects and the other BI tools and how to integrated web results into them!
Just a thought
Matthew
www.logantod.com
Senior Consultant at Martin Wright Associates
04 February 2003 16:02pm
I agreee, an ASP will continue to make ground from the server-side solutions. The main driver is that people are realising that web data is complex, to do it well takes a lot of time and money. An ASP can deliver economies of scale (the development work is spread over their Client base) and reduced risk.
Regarding SpeedTrap my understanding is that it is primarily a usability tool. Good for understanding one visitor's experience at a time.
One of the problems with tracking is that a diverse number of tasks are covered by one name. I belive tracking will diverge into distinct areas of speciality, two of the most important being:
1. Those develivering database solutions (the abilty to improve the targeting of outbound communications based on form data, purchase and viewing history)
2. Those that focus on path analysis and usability (often using agregrate, unpersonalised data)
On 11:20:39 28 January 2003 Ashley wrote:
>Analysts largely agreed last year that web analytics
>solutions offered on an ASP basis were set to overtake the
>server-side (mostly log file analysis) versions which had
>previously dominated (Webtrends being the old stalwart
>most people know).
>
>For example:
>Have a look at this article = http://www.aspnews.com/news/-
>article/0,2350,4191_1183081,00.html about this Aberdeen
>report - http://www.aberdeen.com/ab_company/hottopics/weba-
>nalytics2002/default.htm (table of contents gives more
>detail)
>
>These ASP solutions, provided by people like Clickstream,
>Intellitracker, Redsherrif, WebtraffIQ, Webtrends NetIQ,
>Digimine, Manticore, Nedstat and Webside Story, use
>"browser-based" or "client-side"
>technologies ranging from simple remote calls to another
>server for an image through javascript to java code
>inserts. The data is recorded to the ASP's servers and
>access to reporting and analysis provided by a web
>interface to those servers.
>
>2 things I'd have to say on the above:
>
>1. Do people agree that ASP solutions will emerge as the
>way to do web measurement and analysis? I think they will
>as they are preferrable for a number of reasons....
>
>2. Anyone used SpeedTrap? They're interesting for a number
>of reasons (video replay of users' visits among others)
>but, with reference to the above, they appear to be a kind
>of hybrid of 'browser-based measurement' - using a java
>applet installed on the user's machine - but NOT an ASP.
>They deliver a server to their customers which collects
>the data and provides the platform for reporting and
>analysis. But this server/infrastructure is then the
>customer's responsibility not them as an ASP.
>
>Interested to hear any views.
>
>Ashley
CEO at Econsultancy
04 February 2003 19:49pm
Hi Andy
Yes, I talked to Touch Clarity a year or two ago. What I remember as unique about their proposition was the ability to do real-time (sales) optimisation of web sites. They break the page up into various different ‘containers’ of content and then monitor which arrangement of content is most effective at driving sales. The system can then automatically optimise the page over time – one configuration might work better in office hours and another at the weekend for example.
It seemed to me this would work best for key pages within the site e.g. homepage, registration pages, checkout processes etc. I guess the concern from a site owner’s point of view is – am I prepared to outsource these mission-critical pages?
CEO at Econsultancy
04 February 2003 19:49pm
Hi Ian
You make some very valid points. On point 1, it is true that log files are often the best place to start in order to quickly get some usable data, and, as you say, have access to historical data which is not possible with the ASP approach. I guess I’d still say they were a good place to start for a ‘quick and dirty’ piece of insight but quite possibly not the best longer term solution. A dual offering, as you have, makes absolute sense, particularly if you do want to migrate from one to the other.
Your point 2 is valid, though I have found recently that companies seem to be relaxing somewhat on their outsourcing fears. Transactional data and customer data are, understandably, very carefully monitored and guarded. Site activity data companies are less precious about now. With the advent of web services and the like I think this trend will continue – people will get more used to (and more comfortable with) outsourcing services where there is a good SLA in place and the firm is reputable, has a good track record, good clients etc.
I understand your point about the sophisticated guys wanting to keep control of the data in order to integrate with other sources - typically into a data warehouse for subsequent reporting and analysis etc. However, all the ASP solutions that I know of offer the raw data back to the client if they want it (as CSV, Excel etc.). I know of several instances where the more sophisticated organisations you mention use the ASP basically as a way to collect and ‘clean’ data for them. They see the ASP data as more accurate and more usable and more structured than log files (rightly so I think), so easier to integrate with their other systems. Although they buy the reporting and analysis functions with the service that’s not actually where the value lies for them.
CEO at Econsultancy
04 February 2003 19:50pm
Hi Martin
It’s true that SpeedTrap seemed to develop a name for themselves as a usability tool. I suspect that is because of their mouse movement tracking reporting (uses a java applet client-side), their ‘how much of a form was filled in’ reporting as well as their feature to ‘replay’ a user’s visit. These features appeal to site designers and, indeed, anyone who finds it easier to look at pictures rather than grapple with data (which includes a lot of us, I suspect). Clicktracks (http://www.clicktracks.com) is also quite appealing and unique in the way it graphically shows where users have clicked on a page and you can browse the site like a real user and see the stats as an overlay.
That said, SpeedTrap does a lot more than this. It does all the usual stuff you’d expect. So it should do as the price tag is £10k+ putting it into competition with people like Site Intelligence and Clickstream whose proposition is more about more advanced and integrated reporting – going beyond basic site activity analysis into transactional and user profile data.
Your categorisation is interesting. I suspect you are right that there will be point solutions which essentially cater to different buyers – the database marketers want tools to do segmentation and outbound targeted marketing; the site developers and designers want tools to understand how they can optimise the site; the commercial guys want reporting to tell them about sales, conversion rates and so on. In some case a point solution will be sensible, in others a solution that can cover all elements.
Marketing Director at WebAbacus
05 February 2003 00:18am
Ashley,
It's certainly true that any self-respecting ASP will offer data download, though the data volumes can be very large, particularly when delivered as a CSV or Excel file. And whilst organisations are increasingly bringing clickstream data into their data warehouses, and using BI and analytical CRM tools to analyse this data, online behaviour analysis remains a specialised discipline, where specialist tools are needed. So really the reporting needs to remain in the analysis tool, and the extra data (such as customer records, or CMS data) needs to be imported in on demand. Apart from anything else, using an outsourced model for this would necessitate live database links over the Internet, which, whilst perfectly techincally feasible, is usually a bridge too far for most of our customers.
Cheers,
Ian
www.webabacus.com
CEO at Econsultancy
05 February 2003 09:51am
Hi Matthew
You are of course correct that what is most important is measuring how the web site, as *a* channel to market among others in most cases, is delivering value to the overall business. In most of the cases I've worked on the fundamental questions that (e-business) managers need to answer is 'what value are my web operations delivering? how can we confidently report and measure this? how do we optimise performance?
I agree that it would be helpful to understand 'traditional' BI tools which are of course more than capable of integrating data from web operations to perform analysis - I've done similar work with SPSS in the past. However, I'd add a few notes:
1. As Ian righly points out in his post at http://www.e-consultancy.com/forum/default.asp?v=1527&p=1 "online behaviour analysis remains a specialised discipline" so you still need the right skills and tools to measure the value of the digital channels. Having the data to analyse is one thing. Creating 'actionable intelligence' from it requires a deep understanding of the way the digital channels work - technically, for the customer, for the organisation etc. etc.
2. BI type analysis is one way to approach understanding value delivered. However, it misses out a whole load of other approaches which are perhaps 'softer' such as brand impact, awareness levels, customer experience and so on which need to be measured in other ways - focus groups, surveys, usability testing and so on. Again online can learn a lot here from offline but there are also specialist skills required to do the online stuff (be it online surveys, usability testing etc.)
3. Fortunately businesses are now beginning to accept that web sites deliver value to offline operations and cannot be judged solely on the direct (financial) performance of the channel in isolation. Web sites drive retail store footfall, web sites play an important part in the decision to buy a car etc. even if the transactions occur offline. One financial services provider I've worked with did a lot of business through the web site but 85% was transacted by phone following web research. They knew this simply by having a separate telephone number on the site. I think the whole area of measuring online-to-offline value creation is a very important one. It can also be difficult to do.
So I think there is nothing wrong with having a channel focus (digital channels) as there are specialist skills required. However, the really good people will appreciate how they fit into the wider business context.