Trying out WebTrends and starting to hate it.

Is it crazy to tell my client to reconsider their purchase?

[jay5r]

A client of mine wanted more than I could give them with the AWStats that I had set up (revenue by keyword and referrer, clickpath analysis, etc). Up to this point I've only had much interaction with AWStats and Google Analytics. They decided against Google Analytics becuause they didn't want to give Google access to their data and they have a pre-existing business relationship with WebTrends (the client is a large multi-national corporation). After a little consideration I recommended WebTrends On Demand over the software package because they didn't have proper internal support for the software package (other business units support it, but not theirs).

WebTrends set up a 30 day demo and I've set things up and am starting to look at the data. So far I'm hating the whole experience...

<rant>
For starters it seems way more complicated than it needs to be... It's a case of customization overload - everything can be configured and unfortunately that means everything needs to be configured. I'm seeing options for reports they can't possibly give me results for - things like kilobytes transferred and time to serve pages - they need a log file for those things and we're signed up for the on-line service.

So it's complicated - no problem - I'll read the documentation... Problem is the documentation sucks. If it were well designed you shouldn't need to use the documention very often, but that's not the case. When you do drop down to the documentation you'll find a teaser of the information you need but not the full explanation - it will say something like "More information is available in such-and-such manual" but that manual is nowhere to be found. I then have to start hunting on Google to find the information I need... That's just wrong.

And it's slow. 15 seconds is a typical load time for a reports page, sometimes it's over 20 seconds. One reason it's slow is because they're using graphs for pointless purposes - places where a table would be more informative.

And to top it all off it's expensive - the quote (with a 40% discount) was nearly $10,000...
</rant>

All I can say is that I'm hugely disappointed - I expected a lot more (actually "Less is More" would be an appropriate way to look at it).

I'm really tempted to tell my client to NOT buy WebTrends. As I see it there are a couple of solutions...

First, they could get over their issues with Google Analytics and use it. It's faster, a lot easier to set up, the reports are more to the point (I like the emphasis on revenue-based metrics), and I think it covers what they see as being missing in AWStats.

The other option is for me to write a custom log file analyzer that handles their critical needs. They really only need a couple of things (revenue by referrer, revenue by organic keyword, and clickpath analysis which is not well defined at this point) - yeah they're a bit complex, but it can be done...

And the last option would be to purchase Urchin software. This is the option I feel least confident in, since I've never actually used Urchin software - the closest I've come to it is Google Analytics.

Any suggestions on what I should tell my client? Will I come to love WebTrends once I get used to it, or is my experience typical?

[plumsauce]

Sorry to hear about your troubles.

If you are on Webtrends Live, isn't that charged on a per hit basis? Or is the 10k for a specified level of traffic?

Maybe you can turn off the graphs. That would cut the time that the reporting server spends generating the graphics.

Or maybe their db procs are inefficient. When I slice and dice revenue sources, the report data has sub-second response time at the network card. Speed of transport is another matter since I am on dialup.

Consider this to be a bumping hand since I would like to hear what others have to say about it.

[jay5r]

Thanks for the reply. Yes, the price was for a given level of traffic and features - it is priced per hit.

We're on WebTrends OnDemand - which I think is the newer version of WebTrends Live.

I've got a fast connection to the Internet (cable), so any slowness is on their side.

[cgrantski]

You may have been oversold. To get the revenue information, with referrers etc, you may be able to severely cut back on the number of tagged pages and still get the data you want. The clickpath information is another matter (by definition requiring all pages to be tagged) ... but if the client is just interested in clickpath without really knowing why they want it, I humbly suggest they should skip it and save the money. Or, alternatively, tag all the pages for a month (or whatever), to get a snapshot of clickstream data, then remove most tags to go back to getting only the revenue stats.

Another possibility is to use On Demand just for the revenue/referrer reports, with barebones page tagging. Then, for the clickpath analysis which can get complicated and needs iterative analysis and experiments with filters and such, you can deliver a month of logs to somebody with a WebTrends software license who can dive deep and get you the subtleties and insights that clickstream analysis can deliver if done right.

Or you can move the whole thing to somebody who does WebTrends hosting and who's willing to work on streamlining the data collection under the terms of their license.

My overall suggestion is that you find a good WebTrends consultant (or general analytics consultant if you wish) and give them enough details to get an analytics plan that will give them the best WebTrends (or whatever) solution for their needs at the least cost.

I also think it might be possible to get everything you need from something free or really cheap, and you should continue investigating those. WebTrends' strength is its customizability and if you don't need it you maybe should go with something simpler.

Disclosure: We do WT consulting. I'm not trying to get business, and won't mention my company's name. I'm just trying to point out that this is what WT consultants do.

[jay5r]

Thanks for the response. It just seems weird that a specialized consultant is needed to set up something when an average techie (like myself) could set up things with any of their competitors' products.

Our site is dynamic and two page templates get over 90% of the traffic, so we can't really cut back on how many pages are tagged, but WebTrends was a bit generous with the estimated page count for the year. My goal is to cut back on the traffic without affecting sales...

As I've thought about it (since posting the original message), I've realized I can determine the referring source for purchasing customers myself (with just a little work on my end). I'll talk to them in a few weeks about what they need as far as clickpath information... It does need to be better defined before they move foward.

Bottom line, I don't think we need WebTrends. I think we could do just fine on our own with free stuff + stuff I'd develop, or with something like IndexTools which is reportedly a lot less expensive than WebTrends and seems to have the features we need.

[cgrantski]

Sounds good. And I didn't mean you really need a consultant, only that a consultant who's dealt with this before can probably point out some strategic directions that may not have occurred to you, with just a little more information about the details of your situation. You shouldn't need a consultant to get those revenue reports running at all, and if you are having problems, I too would blame the documentation. It's far better than it once was ... I do have to say that on their behalf.

[martekbiz]

"Thanks for the response. It just seems weird that a specialized consultant is needed to set up something when an average techie (like myself) could set up things with any of their competitors' products. "

There you have it.

You are a techie. I am a marketer.

As a marketer, I know and understand what key metrics and reporting I need to properly assess, analyze and provide post mortums on campaign objectives for my client to help them make better and more informed marketing and sales decisions.

You are a techie. Your job it to implement the page tagging and setup the software (not required with On Demand). Its the marketer who has to provide you with the relevant information that need tracked and logged.

Don't do what you aren't capable of doing. If you aren't 100% familar with web analytics as it relates to marketing and campaign analysis then outsource or have the client hire a consultant for that.

[jay5r]

Don't do what you aren't capable of doing.

Actually, I'm more of a technical business analyst / applications architect than a straight techie - I've never done well in positions where I was just supposed to be "the techie". The fact of the matter is I'm the one that's mostly pushing them on the need for better analysis. Yes, I wish they had a top gun marketing person on their team, and I don't pretend to be one. But my point is when a competant techie can't figure out the technical details without specialized training, then something's wrong. We're not talking about learning a new programming language - we're talking about setting up web site analytics.

I understand all the basic concepts and could set up any of their major competitors' services, but WebTrends is just absurdly complicated which is compounded by horrid documentation. I'm sure it would make sense once I spent enough time with it, but it's a sign of a poorly implemented product when things are this complicated. I remember 18 months ago comparing Urchin and WebTrends and thinking WebTrends seemed really scattered, where Urchin was clear and focused - this is their sales and marketing materials I'm talking about, not their implementations. Google Analytics (Urchin) is still clear and focused and surprise, surprise, WebTrends is scattered and confusing. It's like a lack of clarity permeates the whole company...

[cgrantski]

I have to agree for the most part. It's always amazing to me 1) how hard it is to get over the learning hump with this program and 2) how easy it is for an experienced person to get other people over the learning hump by just adding some explanations and a little hands-on demonstration. I've had to write my own documentation. Ugh.

WebTrends in recent years has chosen a slightly different path from the other products and it has both hurt and helped them. The product is definitely more complicated to use than the others and also very powerful because it puts a lot of customizability into the users' hands. I can't count the number of times I've heard frustration from users of other programs about information they can't get at all, which could be done in WebTrends by an advanced user. WebTrends has done a rotten job of communicating that, and of getting people quickly into at least the first couple levels of complexity. They have also made the interface more complicated than it actually needs to be and have thrown some things into the "customize this" bin that should be in the "works out of the box" bin.

[jay5r]

cgrantski - You should think about doing a WebTrends tutorial for something like lynda.com... If there are that many people frustrated by it, then you've got a market and can set yourself up as an expert/trainer...

I had a conference call with the client and it looks like they'll pass on WebTrends for now. More than anything the price just can't be justified at this point. I'm sure they'll come back to it in a couple of years when the site is more complicated, but for now it's overkill.

[Tineke28]

Hi

I use intellitracker and find it's interface is much easier to set up and use from a marketers perspective. Also, all the webbie needs to do is put page tags on and Intellitracker does the rest.

I used webtrends and it was a nightmare - too techy, inaccurate information, etc.

By the way, I don't work for Intellitracker, I have used it at two jobs and found it to be really good.

Thanks
Tineke

[martekbiz]

"inaccurate information"

Obviously you aren't using the new version of web trends with their inpage tagging.

The information you get is the information YOU set it up to track and filter.

[cgrantski]

Too bad about that. A lot of people try a lot of products, find them unsatisfactory, and assume that it's the product's fault and that, years later, the product hasn't changed either.

Same with politics.

[Pibs]

I can well understand someone not wanting to hand Google all their conversion and business data.

The temptation to charge a little higher for keywords if using Adwords for example, seeing as how obscure the system is, would be very high.

You are essentially telling Google just how much you can afford to pay for advertising. That strikes me as kindda dumb?

If you were thinking of buying a used car would you tell the salesman exactly how much you had spare in the bank? I cannot really think of any situation where you'd want to pass on data to your advertising company regarding just how much money you're making.

One option for a basic system is Statscounter, which is freebies for a small account. Doesn't do conversion tracking I don't think but can certainly pinpoint paths and sales funnels etc.

Awstats is (virtually) useless.

P.