| 3:05 pm on Aug 23, 2004 (gmt 0)|
There are TONS of excellent solutions to solve your exact needs. I would say that the best for your situation would be a 3rd party tracking system. These usually work by adding a small snipped of code to all of your pages.
From there you can track just about anything and everything you'd ever want to know about a visitor. If you have a few bucks, you can even get a solution that will automatically pull your cost per click from both google adwords and overture. A true marketers dream come true!
Let me know what you are hoping to track, and how much you are willing to spend on this tracking. Decent ones start at $15 per month and others in the $$$thousands. Depending the depth of cost per clicking is the money issue. If you need in depth stats, then you can do this pretty cheap.
Rather than having to look through rough log files and lots of junk, these solutions simply make it organized and searchable. Definitely worth the money.
Let me know if you need help, or sticky me.
| 3:13 pm on Aug 23, 2004 (gmt 0)|
I've got 2 SEO contracts coming up so It would have to be a reasonably good value solution, but would be good if it were "certified" and not hosted by either me or the client so it could be deemed "independent".
I need to track where users came from and what search terms they used so that I can furtehr optimise the site. I will be billing monthly based on cost per click per keyword so will need to know exactly how many unique visitors we've had. Ideally if they come back, we recognise them too. As logfiles are just text and *could* be tampered with, I don't think its fair to base it on that. We could be talking £1000's of pounds a month and both want to know figures are accurate (especially if these are not turning into sales!)
I've been looking at Deep Metrix and also Web Trends.
Any recommendations appreciated.
| 4:09 pm on Aug 24, 2004 (gmt 0)|
I've done this arrangement before. However, I would raise caution with your statment that you are billing based on the cost per keyword.
Are you really intending on breaking out every single search term and putting a cost to it? Just the idea of updating the cost alone on a monthly or weekly basis makes my head swim.
Furthermore, I doubt you are going to get an off the shelf solution that will list every single keyword that was used to reach the site. Typically they only display the top 20 or 50 or so.
I would suggest taking a weighted average of a sampling of keywords (high volume and low volume, expensive and inexpensive) to get an approximate average cost per visitor. Then just multiple this by how many visitors you brought in.
A simple report of the top keywords used by visitors should be enough to validate that your approximated calculations are still on target, thus validating your pricing.
| 10:24 pm on Aug 24, 2004 (gmt 0)|
If I were a client I wouldn't want to pay by the click, what counts is sales and profits.
It's incredibly easy to get loads of clicks, generating profits is generally much harder.
| 11:07 pm on Aug 24, 2004 (gmt 0)|
If I can get the customer to his door, then the rest is pretty much up to him: having a compelling product at a competitive price. I can help him make his site as usable as possible and get as many prospective customers to his door as possible, but if they don't buy the goods, there's not much more I can do. Got any solutions?
| 11:13 pm on Aug 24, 2004 (gmt 0)|
>> I would suggest taking a weighted average of a sampling of keywords (high volume and low volume, expensive and inexpensive) to get an approximate average cost per visitor. Then just multiple this by how many visitors you brought in.
I agree: the approach I'm planning to take is to take the average AdWords/Sitematch value and halve it. I don't want a mainetnance nightmare. Maybe I should do this monthly or every X months?
| 11:53 am on Aug 25, 2004 (gmt 0)|
I agree with geebee2, from the clients point of view clicks are just an expense, what they are really interested in is the sales.
Have you thought about a commision based deal? If its possible thats what I usually go for.
| 11:57 am on Aug 25, 2004 (gmt 0)|
P.S. In regards to stats, I have heard that clicktracks.com is good. But i haven't used it my self - there is a free trial....
| 12:07 pm on Aug 25, 2004 (gmt 0)|
Having used this model for some time, I have to say that clients love it IF (and it is a big if) you get the stats right.
Poor stats mean arguments with the client - so, in the end, we developed our own stats package and agency/client interface allowing clients to remove phrases they don't like from billing, giving good word overrides, installing abuse filtering etc., etc.
Not a lot of use to you as we don't sell it to third parties but it does indicate that the model can work very well indeed for both the SEM and the client, provided both parties have full confidence in the system.
>It's incredibly easy to get loads of clicks..
I think what we are referring to here is sending highly relevant traffic - a little more sophisticated than just sending click-throughs.
| 4:43 pm on Aug 25, 2004 (gmt 0)|
I have discussed commission/profit share but he is used to paying for clicks and actually wants to. Only time will really tell whetehr this will work business-wise. As far as tracking sales, I guess osme sort of affilaite tracking engine could do this, and would need cookies etc. Interested in what approach people would recommend for this (or a pointer in the right direction)
What I need for the CPC, as quite rightly highlighted, is a third party stats counter that will filter spoofed/automated clicks, certain IP addresses (such as mine and customers) and generally give us both piece of mind that we trust the figures.
Thanks for the refs to date. Still interested in more recommendations and reviews of webtrends or deepmetrix for these purposes.
| 8:25 pm on Aug 25, 2004 (gmt 0)|
Again, just a clarification that might help you achieve a better price for yourself. You mentioned you would take the average of the PPC price. However, I still suggest a weighted average.
The reason being that most of the higher priced terms are also the higher volume terms. If you simply average those bids with lower ones, it will be lower than the Cost per Click that Adwords or Overture would give you. Simple example:
200 searches for "widgets A" @ $1.00 / click
50 searches for "widgets b" @ @0.10 / click
If you average those bids, you would get $0.55 as your price per visitor.
Whereas if you take the weighted average, you would have [(200 * 1) + (50 * 0.1)] / 250...$0.82 per click.
If you assume the client bid on those, and achieved all the clicks, that is exactly what they would pay - $0.82 per visitor.
I believe that is the most fair way to calculate the number, for both parties. I do agree with the enticement to the client though of offering 25% off that price; effectively making it cheaper than the traffic achieved through PPC.
We've used this numerous times. Prospects who understand it, love it. Prospects who don't understand it...need more help than it might be worth anyway!
| 4:39 pm on Sep 6, 2004 (gmt 0)|
Thanks for this it makes a lot of sense.
One final one: how best to find what Google/Overture sell these keywords for. Do I have to go in manually each month and bid to get the values? Or is there a summary of what they were worth over the last month period?