Web analytics targets : what's good?
Image by Bain News Service,, publisher.
I actually think its pretty easy to measure things on the web. There are a million tricks of the trade to allow us to capture just about any actions on a website and correlate it to campaigns, segments of visitors or site functions. Even figuring out what to measure isn't that complicated once you corner someone and get a straight answer on why a campaign or site exists. The hardest thing for me is determining whether the measurement is actually good or not and what a decent target should be. Its not a simple answer, especially if you are starting from scratch.
In the last couple of days, this topic has come up in multiple conversations about clients and setting up goals and targets for new websites and campaigns. The paraphrased questions were as follows:
"We have 15,000 signups. Is that good?"
"We have time on site and page views per visit metrics for the site today, can we see how they fare against industry benchmarks?"
"We haven't been measuring how many people get to the product catalog prior to the redesign, so how do we set a goal?"
I don't have a standard answer for questions like those, except my fall back answer about benchmarking is that I don't think its possible to compare your site to any other site with any sort of validity. Even if you are in the same industry there a million other factors driving things like conversion rate, time on site, and page views per visit. In most cases you don't know how those metrics are even being calculated or in what context. Example: Site A might average 10 minutes on site while Site B might average 8 minutes on site? What's better? What if I told you Site A has a confusing as hell navigation that drives people to click around desperately trying to find something and has negative customer satisfaction ratings based on the abysmal navigation? All of a sudden, having a high time on site isn't looking so rosy.
So now that I've rained on the parade of benchmarking, are there more constructive answers to these problems?
My feeling on what is good or bad still revolves around benchmarking but only in the case of benchmarking against your own companies' performance. My mantra is of continuous improvement, not comparing your web metrics to your competitors or industry (with the exception of audience size, keywords, and sources of traffic). Let me revisit the earlier questions with ideas on how to attack those.
"We have 15,000 signups. Is that good?". I actually don't know if 15,000 is good or not, but I would rephrase the way we are asking the question. Probably a better way to think of this is how much did the 15,000 signups cost us? Lets say the 15,000 signups are a result of spending $1.5M in tactics to drive people to the site. You now have a metric of $100 cost per acquisition. How does that stack up to previous efforts to acquire names for your company? Is that better or worse? Another question to ask is what is the value of a signup? Is the lifetime value of a signup more than $100? If so, you're probably doing alright.
"We have time on site and page views per visit metrics for the site today. Can we see how they fare against industry benchmarks?" As I mentioned earlier, its very difficult to use industry benchmarks with any kind of certainty. Every site is different, regardless if they are in the same industry or not. I'd rather focus my time on improving what is going on with my site. Let's say time on site is currently 5 minutes per visit. We devise some new functionality on the site to get people to explore more content and stick around longer because we've determined that folks that spend a lot of time end up as leads. Measure the uplift (we hope) of those efforts in driving that goal. Keep improving by using your own site as the benchmark. Do landing page optimization, multivariate testing, segmentation etc to drive improvements.
"We haven't been measuring how many people get to the product catalog prior to the redesign, so how do we set a goal?" Ah, the trickiest of the lot. What do you do if you have no baseline whatsoever? Basically anything you get is an improvement since you are starting at zero, right? This situation is hard to have a target to throw into a spreadsheet and measure against. Instead what I would suggest is measure product catalog visits as a goal and see what you get the first month or so. Then set that as the baseline and work on ways to get more people in the catalog while measuring the effectiveness of those efforts. See what activities and traffic sources are driving the product catalog, optimize accordingly.
I don't believe I have the definitive answer for creating targets, but these were just some of the ways I handle the problem. Does anyone else out there have other strategies to share on creating targets? I'd love to hear it.
Comments
No comments yet. Be the first.
Leave a Comment