All Web Analytics Applications Should Be Free! Not.

RosesA dear friend who works for a nice Web Analytics vendor wrote in an email something that surprised me a little bit. If someone else had said that it would not have been such a surprise, but it was someone whom I know well and for whom I have the greatest respect.

In the email the statement that my friend made was:

[ … ] your repeated insistence that web analytics applications should be free [ … ]

 This really made me think along two angles:

    1] Hmm… does not sound like I have ever said that explicitly or implicitly.

    2] Do I really think that? Should all applications be free? Can they be free?

A few moments of introspection yielded a series of beliefs I hold and have often stated on this blog or during speaking engagements:

  • Spending on web analytics applications should be a fraction of web decision making budgets: Yes.

  • We should take spending away from Vendors and pour it into people to analyze the data. Yes.

  • You should not jump in and buy Visual Sciences or Omniture or ABC as your first web analytics implementation. Yes.

  • You should know that clickstream is just 30 – 40% of data you need to drive insights. Yes.

  • 98% of the clients of existing web analytics vendors (all of them from Omniture to HBX to Google Analytics to ClickTracks to WebTrends and on and on) are not at a level where they could be a solid case study for their vendor because they simply can't show impact from web analytics (and not because they don't want to do a case study). Yes.

  • Most people out there don't need one of the big three vendors (but still have them). Yes.

  • A tiny fraction of companies in the world, the ones with huge and deep pockets, can exploit the big three and expensive vendors. Yes.

So I fess up to all of the above, a big yes, but my conclusion was that I have never said all web analytics vendors should offer their applications for free. Hopefully you agree with my self diagnosis, if not please share your feedback via comments.

[I certainly believe that in the long run all of the above truths are good for the industry and web analytics vendors because the output (if anyone follows any of the above recommendations) will be smarter clients and analysts who can actually use all the features the big tools possess and smarter client who will push vendors to innovate and push the analysis envelope (in turn helping web analytics vendors to charge their clients more). Win- Win-Win.]

That lead to the second part of the thought process, should all web analytics applications be free? Can they be free?

After more moments of introspection the answer was No. I don’t believe that all web analytics applications should be free or that they can afford to be free at the moment. Reasons include:

  • Everyone has to eat, including the analytics vendors !! : ) More seriously….
  • Free usually works, or is an imperative, in commodity businesses. Web analytics is not there yet, and neither is the “web”.  Maybe basic log file analysis is a commodity, but there is little in terms of actionable insight that one can get from that.


    Web Analytics is a baby, it has a lot of growing up to do. Some day it will be all grown up and mature and a stable.

  • Perhaps most importantly the world wide web is evolving at amazing speed. New interaction models, different ways of contributing, constant redefinition of content and on and on. Given all these challenges we need the existence of for profit enterprises who can pour resources and $$$ into adapting to all these changes and helping all of us measure success of our web efforts.


    Open source is usually sub-optimal in such scenarios (simply don’t have a large enough group out there for web analytics and they don’t have the kind of time or money to keep pace with change).

  • Some people make money selling razors, others make money selling razor blades, still others make money coming to your house and giving you a shave (precluding the need for you to buy either razors or razor blades).


    Most Web Analytics vendors currently don’t have business models that are much more beyond selling software and some consulting services. They aren’t diversified enough to try different things. So they can’t afford to go free (all except for the obvious one).

Hence atleast for now, except for one, most web analytics applications won’t be free and can’t be free. For now that is just fine for us the customers of these applications because we need for profit enterprises to do lots and lots of good research and development to keep pace with the rapidly evolving web and help us measure what we need so that our brilliant analysts can find insights which ultimately will benefit our customers. Circle of life.

But this does not mean we won’t see some more free versions in the market soon to counter the “Google Analytics effect”. Atleast to protect the “bottom” and middle of the market it is quite possible that we will see a free version of Omniture. Basic metrics and 100k page views per month for free and then a tiered structure pricing. This protects against Google and, soon to arrive, Microsoft’s massive sucking up customers and allows Omniture to make this customer friendly statement: “we’ll hold your hand and walk you to our paid services when you are ready because we have some very complex and sophisticated solutions for you”. Omniture wins and Goolge Analytics & Microsoft have to fight harder and customers win. Cute for us.

This is not too radical a thought, a simple extrapolation from the lessons The Innovator's Dilemma teaches. Mr. Christensen's excellent book talks about technology as a disrupter, in this case the "free high end analytics" is the disrupter.

[IMPORTANT: Omniture is used just as an example since they are industry leaders, you could swap them out for any other major vendor. I have absolutely positively no inside information and have not spoken to anyone at Omniture. What is stated is pure hypothesis, or if it makes you happier, a conspiracy theory.]

In closing, do take a critical look at the web analytics solution that you have and if you have implemented the 10/90 rule in your company. If you have not then it is highly recommended that you fix the balance to be in favor of your analysts. Grow from Free to Expensive and along the way push each vendor you have to provide the value that they promise they will in the demos.

Do you think all web analytics applications should be free? Do you think they will all be free in the next decade? Any other relevant factors that might influence the vendor pricing decisions? Please share your feedback via comments.

[Like this post? For more posts like this please click here.]

Comments

  1. 1

    I'll try really hard to not evangelise. :-)

    "…applications should be free"
    Is a commonly misunderstood myth of Open Source and/or Free Software. Think freedom, not zero cost. There is a huge difference. I personally make a far-more than respectable living from Free and Open Source Software.

    "Maybe basic log file analysis is a commodity, but there is little in terms of actionable insight that one can get from that."
    Then you're not trying hard enough!
    There's plenty of *very* actionable insight in basic log analysis. Perhaps it's not as easy to tease to the surface, but that's more the fault of the tools than the data.
    My polite disagreement Avinash, is more around your exposure to the high end tools and having highly tuned websites. vs most people are still in the "I'm getting X hits a day" phase, *any* insights are gold. You're at the wrong end of this stick. :-)
    The insights are there, I extract and use them to great effect every week.

    "Open source is usually sub-optimal in such scenarios…"
    Avinash, this very website of yours is Apache on Linux running WordPress, which is written in PHP and has a MySQL backend. All Open Source, and mostly GPL'd (ie Free Software) at that. Our email exchanges we've had have only gone via Open Source software. :-)
    The reality is much of this very fast pace of change is driven on or by Open Source and Free Software. Ironically…

    "…we need for profit enterprises to do lots and lots of good research and development .."
    You should check out some of the freely available university research papers in this space then. There is stuff I've come across, really seriously clever stuff, that I've not seen in *any* commercial product.

    This is not the place to argue the point in detail, but don't fall for the trap of thinking that all good ideas only come out of commercial software companies – despite who you work for. ;-)
    This very website of yours, I put forward as exhibit A. Why aren't you using commercial products for this website?
    I don't recall there *ever* being a commercial blog software market that was eventually commoditised. Or did I blink and miss it? :-P

    I'm not, even remotely, anti commercial software, it'd be hypocritical for one – esp as I'm a long time fan and user of your very companies software. :-)
    My concern is more around most peoples awareness or even understanding of Open Source and Free Software. I flat 100% guarantee you that *every* visitor to your website has used Open Source or Free Software. And most of 'em won't even know it.

    I don't think all web analytics software should be zero price, and I sincerely doubt it ever will be. The business model of "selling" it may change though.

    your argumentative, but loyal, reader :-)
    – Steve

  2. 2

    Steve: I welcome "argumentative, but loyal, readers"! How else would I learn. :)

    Like yourself I am a huge fan of open source software and have gobs of it on my pc, not just this blog.

    I think we disagree perhaps on couple small things:

    1) The rapid pace at which innovation is required in the web analytics field.

    2) Existence of a core community of "developers" to help us keep pace with the development.

    I'll be thrilled to be proven wrong because if I am wrong and useful WA applications become open source then it will be that much easier for companies to implement my 10/90 rule! :) [And that 90 of course would go to wonderful folks such as yourself!!! :) ]

    Thanks for disagreeing, your thoughts are always welcome and much appreciated.

    -Avinash.

  3. 3

    Well, I'm glad you wrote a post about Visitor Segmentation in the Web Analytics Industry.

    There's a lot of good "free stuff" out there – but won't fly in a large corporate enviornment. Depends on the audience. If 90% of all the people who use Web Analytics have under 100K pageviews a month, the free solution works.

    I personally get a lot out of SiteMeter for WebMetricsGuru.com and have grown to like it alot. I am also likeing, for different reasons, Performancing for my ArtNewYorkcity.com blog. But I would never ask, or expect these products to scale to an enterprise enviornment.

    In other words, I'm a different customer than the corporations and businesses I work for and with.

  4. 4

    Avinash, regarding your points 1 & 2, I don't think we do disagree – maybe in the details. :-) See if I can explain….

    Both are issues – undoubtedly, both do have solutions if we're prepared to go get them.

    Focus on point 2 first: developers.
    How do you build a community of developers? It's hard work, and there's lots written about it. Basically you need to excite them. Hit their drive nerve. Give them something they *want* to work on.
    This is part education. Which I suggest you're doing a superb job of. Demonstrate the problems, encourage people to join in and help.
    It doesn't magically happen on it's own.

    The flip side is that people are doing this for their own reasons. "We" have no demands to make on these people. We can suggest, encourage and provide feedback. Gobs of positive feedback is incredibly effective.

    The other part is assuming that the bulk of the work is done by the developers. This is also a myth. They do the coding – sure, but what about documentation, *testing* and even feature discussions.
    I personally find that getting involved in feature discussions can be hugely rewarding and beneficial.

    You're trying to build a community, not a group of slave labour coders. If you approach it as the former, everyone wins. Approach as the latter….

    And this is the key point:- you're building a community, some who just so happen to be developers, of which a smaller number will be sufficiently engaged to actively code for the community as a whole, and not just submit the odd patch.

    Now to the first point. If you've got the 2nd, the first will follow. And in my personal experience, innovation can and will move so fast your head will spin.

    And that brings us back to the educating.

    WA as a field could be increasing it's pie size by gradually educating all 98 million websites (Netcraft Oct 2006 Survey) that WA can be used to good effect. And demonstrate how easy it is to do so. Demonstrate what the problem set is. etc etc etc.
    Help and Encourage the website owners to actively fix problems with their sites and make their sites work better. The rest will happily follow.

    @ WMG:
    I understand what you're saying and even agree with the vibe, but not the secondary statement. :-)
    I *only* use "free stuff" on sites that handle several million pages a month.
    I've tried several commercial products – both bought and paid for (several $thousands bought and paid for…) and demos and generally they're junk. Painful, slow, horrible for useful insights etc etc etc. And don't even get me started on their customer support. Or rather lack thereof.

    As Avinash mentioned in this article referring one of his earlier ones – it's not the tools, it's the people.
    Ansel Adams was able to take the most amazing landscapes with tools that would be regarded as horribly primitive by todays photography standards. Didn't seem to interfere with the quality of what he produced! Glass Plate Film!!!

    "Depends on the Audience" I do agree with. :-)
    Sure there are things the high end tools do that are amazing. But if you don't use those features who cares? :-)
    GA was pleased that they had 1/2 a million customers a while back.
    That's what? about half of a percent of all web sites? Plenty of room for growth!
    The high end tools are for such a tiny portion of the market. Way down the far end of the magical tail. Imagine how big the market could be if those 97.5 million websites all realised just how useful WA could be to their sites???

    Where I feel the current crop of tools, and I include *all* tools in this claim, fail so very well is that they really don't help Mrs Average. It's all jargonised and technical and mind numbingly obscure.
    Where's the simple half page reports with simple *easy* to implement suggestions on ways to improve Mrs A's website. Simple Suggestions that *WORK*.

    We as professionals in the field can do this, why can't our tools do it for us?

    And no, I don't think "website" clippy would be the way to do it. ;-)

    Cheers!
    – Steve
    PS. Sorry 'bout the length – just got on a bit of a roll there… :-)

  5. 5

    Avinash,

    Excellent post as usual. Since we'll be able to hash this out next week in person, I have a few questions. ;-)

    You've repeatedly said "We should take spending away from Vendors and pour it into people to analyze the data" which in-and-of-itself I don't disagree with. I've been saying "Staff. Now." for a few years and am encouraged by the progress we're seeing as an industry to be sure.

    You also have said "You should know that clickstream is just 30 – 40% of data you need to drive insights" which I also don't disagree with. I'm less sure whether the balance of the data is simply "VOC" and "outcomes" or something more nuanced and complex, but yes, clickstream is not the whole problem.

    Given these two valuable insights, and given Intuit's demonstrated ability to hire some very bright people, can you/have you blogged about how companies should go about spending that "90%" dollar amount? Meaning, how should they actually go about hiring people and where should they look?

    I ask this because I'm constantly torn by your arguement that the majority of investment in web analytics should be in people. I mean, these people will need books on analytics and I'm very much in favor of that, but what if the people simply aren't available? What if they're too expensive to justify? What if your company isn't as dedicated to the web analytics business process as it pretends to be and the talent you hire bails in 90 days? What if you don't have the hiring resources of Intuit?

    Don't get me wrong. I love the recommendation. I'm just not entirely sure it's really that practical. Personally I get an awful lot of email from companies trying to hire folks and the available pool of talent actively seeking jobs just doesn't feel that deep to me right now. This is aggrevated by the recent wave of folks who have decided to move from non-vendor companies into the vendor-sphere, shrinking the pool at a time it needs to be expanded.

    So two questions back to you:

    1) For the companies that want to hire dedicated resources, where should they look?

    2) For the companies that are unable to hire, do you recommend either A) going with a for-fee vendor that provides implementation, business and strategic support, B) that they keep looking and do the best they can, or C) something else?

    As usual I look forward to your response and the ensuing conversation. Hopefully we'll be able to continue the conversation in D.C. next week.

    Your loyal reader,

    Eric T. Peterson
    Author, Web Analytics Demystified
    http://www.webanalyticsdemystified.com/weblog

  6. 6

    Avinash,

    Great post. I agree with you entirely that not all web analytics apps should be free (and I speak as someone who works for a company that will be launching a free web analytics app!).

    In the web analytics industry, there is room for 'free' and paid-for solutions. The 'free' solutions will mainly be subsidised by other money that its vendor can earn – in the case of Google & MSN, this is of course advertising revenues. But nobody should be under the illusion that Google (or us) is providing web analytics out of the kindness of their hearts – and the price you pay is that you're entrusting your data to a non-impartial player.

    I'm not suggesting for a moment that Google would do something evil with the data they're entrusted with, and we're having the same kinds of discussions here about reassuring people that despite the fact that we have our own agenda to pursue (bringing people to our ad network), we will maintain the highest standards of impartiality, and actively work to help people understand marketing response across Microsoft and non-Microsoft channels.

    But many organizations will want to deal with a truly independent vendor, and, furthermore, one whose product development decisions are not influenced by the desire to monetize their ad network. In that case, the software's got to be paid for somehow, and at least if you're paying for it yourself, you have a nice straightforward relationship with the vendor.

    I'm not sure I would agree with your prediction about Omniture (or WT, or WSS) releasing a free version – the independence argument should be strong enough. I guess the only reason is to seed the market and have some folks on a free version of their own tool (making switching costs lower for a higher-end tool) rather than Google's or our product. But, ironically, I think that people will be keener to start with Google Analytics because they won't fear being upsold aggressively by a vendor who they know has the ultimate goal of taking money off them. It'll be interesting to see how it pans out.

    Cheers,
    Ian

  7. 7

    Avinash, great post! I pretty much agree with you, mainly in this point:

    – Web Analytics is a baby, it has a lot of growing up to do. Some day it will be all grown up and mature and a stable.

    I work everyday with the most well known analytics solutions, and what I feel is that just don't fit completely my expectations, and I fount the answer in two mainly pillars:
    – Flexibility: Since there is a lot to do what I mainly need is flexibility, mean the possibility to build over the standard solution, and I’m not talking about build just reports, I’m taking about the data system that will be processed in order to get meaningful information for every specific project. As you said, Web analytics is a baby and will not grow healthy unless the Analytics Companies do not work together with analytics professionals.
    – Integration: I’m not saying that these applications companies are working focus on integration, even worst, some of the big ones don’t have at least API. So I need to integrate every module of my BI system manually? Internet allows having tracking all the information you want in real time, so why don’t have it integrated?

    I’m working for fortune 500 companies and even so, I guess Analytics companies’ prices are pretty ridiculous.

    Well this is just my point of view, I’d love to receive some responses.

    Juan Damia
    http://www.damia.com.ar

  8. 8

    Eric: It is so easy to reply back to your comments because you use bullets and sub bullets and have such a logical flow. As you might have observed I do that a lot and I really appreciate the structure, thank you. Ok on with the answers, or atleast a stab at them……

    Perhaps I am misunderstanding something here Eric, hence I could be wrong about this, but if this were really true in your company, or for your client:

    but what if the people simply aren't available? What if they’re too expensive to justify? What if your company isn’t as dedicated to the web analytics business process as it pretends to be and the talent you hire bails in 90 days? What if you don’t have the hiring resources of Intuit?

    Why would we spend any money on a web analytics application in this case? If we don't have any one to analyze the data in your company why buy Visual Sciences or WebTrends? The 10/90 rule in the above scenario might be more like free/0 (where free = Ian's upcoming solution from MSFT and 0 = we don't have anyone so let us have the admin send some pdf versions of reports via email).

    But if the above is really true then a great option in spending $90, still holding $10 for the tool, would be to go get "big brains" from outside. Someone like:

    http://www.zaaz.com in the US
    http://www.logantod.com in the UK
    http://www.ox2.be in Belgium

    They can provide the nice analytical horse power required and when we find someone we can hire we can transition in-house.

    I am going to cheat and give brief answers to the other questions Eric, got to get to work….

    1) For the companies that want to hire dedicated resources, where should they look?

    Look in unusual places and then give them Eric Peterson's books :) and teach them. So for example look for smart financial analysts or people in the traditional BI world. Look for two criteria 1) that they are inquisitive and 2) at some level they "get" the web.

    The problem with most hiring processes is that one looks for people with 15 years of experience in a industry that has existed for 5 years. Hard core web analytics experience would be great but is usually overrated (as compared to other skills required to do the job).

    2) For the companies that are unable to hire, do you recommend either A) going with a for-fee vendor that provides implementation, business and strategic support,

    If the company has nothing then go "for free" vendor and yes look outside for analytical help, like the above three url's.

    B) that they keep looking and do the best they can, or

    For companies of some size (small) this might make sense.

    C) something else?

    In no scenario should you go without a tool. Even if you have no analytical horse power and you can't afford to give money to vendors. Get a tool, get google analytics or Ian's MSFT offering.

    It is free, implementing a tag is trivial amount of work and you'll atleast start the process of educating your organization with however little data you can use. It is free! :)

    Thanks again Eric for your wonderful comments, to have a industry luminary such as yourself as a reader and contributor to this blog means a great deal.

    -Avinash.

  9. 9

    Juan, Ian: My comments on your comments on my comments.. :)

    Juan:
    There is a fundamental mindset shifts required for most web analytics vendors. Most if not all believe that they are the center of the company decision making universe and that is almost never true. If a vendor believes they are at the center of the universe they will work on getting data into their tool, which is what most have done. What is needed is a much easier way of getting the core data out to merge with other sources in the company. This is what we have done with our Data Warehouse approach for end to end analytics, but as you put it we have done this "manually". But have not fear I think change is in the air and most vendors are starting to, or will have to, play ball.

    Ian: On the thought of if Omniture or WebTrends et al will offer a free version…..

    You make a fabulous point about the seeding part. I was not even thinking of that. My thought was that GOOG and soon MSFT will clean up the small and mid sized businesses and reduce the number of sites "in play". To counter that some of the current vendors will have to come out with free so they are on the table atleast.

    This is important for two reasons:

    1) Many customers, self included, are "lazy". If some of the free tools "work" then they won't switch. Now I realize that free tools won't make you coffee in the morning along with giving you reports, but there are fewer customers who want that.

    2) I am assuming that what we are seeing from GA is version one of the tool, and what we'll see from you is version 0.9 :). That means you'll get better with age (like fine wine). If you, and GA, do then you are a much bigger threat.

    Important Caveat: I am just a simple web analytics person and not a business strategist and so take my "strategic" analysis with a grain of salt because it is not really my sphere of expertise.

    Thanks for your comments guys, appreciate that very much.

    -Avinash.

  10. 10

    Avinash,

    I am a recent Finance graduate turned web analyst. In response to "so for example look for smart financial analysts or people in the traditional BI world," I must say you are right on:) The strategic thinking and mindset that is learned in a Finance curriculum can successfully be applied to web analytics. Instead of forecasting cash flows, or analyzing balance sheets, I'm crunching abandonment rates and analyzing navigation behavior. It requires the same thought process, except I don't gotta wear a suit to work. 3 months ago I had no idea what a conversion rate was, and now I'm the web analyst for my company. Your site has been a valuable resource and I thank you for that.

    Tommy
    http://www.reddoor.biz

  11. 11

    Hi Avinash,

    I couldn't agree more, at OX2 we are sick and tired of seeing big corporations with top level tools that don't get used. I have a new customer for example that bought WebTrends 4 years ago and the product had never left the box… Since the begining of our Web Analytics activities 3 years ago we have been trying to evangelize our market and communicate the importance of dedicating ressources. This problem was for example addressed last year by Aurélie on our blog post: First Web Analytics (wednes)Day, Brussels.

    When someone asks us where to start we often recoment 'take a Google Analytics, play with it a couple of months and give us a call'. This allows the user to see what kind of information is available and with a little bit of chance thay will come back asking questions. We can then discuss the implementation of a Web Analytics tool (at OX2 we mainly work with WebTrends, Instadia & Google Analytics) and either the delivery of consulting services or a training/knowledge transfer program. We don't always recomend paying tools as sometimes it's not needed or it's too soon to go towards that direction.

    Regarding your little story about Omniture, I really like your thinking. And you're right that a vendor with that strategy might have a chance of wining the race. Let's see who follows your advice ;-)

    As usual a very interesting post. Keep the good work and see you in eMetrics in a few days.

    Kind Regards,

    René

  12. 12

    This one appears to be free: http://www.breadboardbi.com/clickstream.html

  13. 13

    Hello,
    Application should be free as long as big vendors can offer and charge for other services or features.
    I agree with you, vendors should offer up to a certain service for free, and charge for more advance features. This is what some people call the "freemium" model.

    The Google Analytics model is just a value added service for their adwords program. It helps Google sell more ads, and adword users analyce their results – is a win-win situation. The rest of us that use the Google analytics but don

Trackbacks

  1. Should web analytics be free?…

    Yet again there's an interesting post on Avinash's blog about whether web analytics applications should be free. Avinash has been unjustly accused by a friend of his of asserting that all web analytics applications should be free, which he doesn't…

  2. […] The spectrum is one of price, reflecting the ungoing discussion about the issue surrounding product pricing and whether web analytics products should be free. I must admit not having a clear opinion on that and can only express my view as an economist. Using Web Analytics tools right, undeniably assures increased utility for your company. Your company would thus be ready to pay for such a service. The price tag you'll decide to put upon the acquisition of products and hopefully service, should reflect the commitment your company is willing to take while walking the path towards data driven decision making. Also note that data driven decision making is about culture. Not all companies have such a culture in them and upper management will have to work hard to establish, or at the least reinforce, it.  A global Web Analytics strategy, that should unfold into data driven decision making, can be adresses in 2 main ways: decentralised and centralised. […]

Add your Perspective

*