Representation of high accuracy and low precision.

Image via Wikipedia

I constantly counsel people to make their Web site more than just an online brochure--they must measure their conversions online and offline so they know what it's worth to bring that extra visitor to their Web site. Only then do you know what to spend on marketing. Simple, right? Not always. What do you do when you have multiple ways of measuring conversions?

First things first. If you don't know what a conversion is, or how to measure conversions, you need to start there. Whether you sell online or offline, you want to engineer your customer experience so that each sale can be tracked back to your Web site. That way, you can do the math backwards to know what it's worth for each new visitor to come to your site because you know how many visitors completed the last step the Web site offers.

For example, if you sell online, the last step is completing the checkout--simple. But if you sell offline, you might need customers to call a special phone number or bring a coupon to the store or fill out an e-mail contact form--anything that helps you identify them as coming from your Web site. By counting how many visitors come to your site, knowing how many of them fill out the e-mail form, for example, tells you your Web conversion rate. If you know how many of those e-mail leads turn into sales, then you can calculate your overall conversion rate, and what it's worth for each visitor who comes to your site.

So, you might know that 10% of all people that fill out the e-mail form eventually sign a $50,000 consulting agreement with you. So, getting someone to fill out that form (your Web conversion) is worth 10% of $50,000 in revenue, or $5,000. If only 1% of the visits to your site fill out the contact form, then each of those visits is worth 1% of $5,000, or $50 in revenue. Your marketing budget should reflect spending considerably less than $50 a person to get someone new to your site--if your profit margin is 50%, then you're making $25 in profit for each visitor, so every dollar you spend on bringing people to your site cuts into your profit.

But what if you know all this already and you are concerned that different analytics programs seem to tell you a different story each time you try one? Suppose you try Google Analytics and it tells you you had 29 conversions based on 3,011 visits last month but that you also have WebTrends installed and it says that you had 31 conversions based on 2,831 visits?

Who's right?

Well, probably neither one. None of these analytics programs are totally accurate. They each have errors in the way they collect the data. But you shouldn't care which one is correct. Nor should you worry about this level of accuracy. What you should embrace is the trends they show you.

You only need one analytics program. Yes, it won't be accurate, but that's OK. What you want to examine is whether the things you are changing are improving conversions or not. Whatever inaccuracies are inherent in the analytics are stable from month to month, so just pay attention to whether your site's metrics are improving, based on the actions you are taking, or getting worse.

So, rather than being concerned about a standard way of measuring conversions, pick a way that works for you and start making decisions on what the data tells you. The secret isn't accuracy, it's action.

Reblog this post [with Zemanta]

February 11, 2009

Mike is an expert in search marketing, search technology, social media, publishing, text analytics, and web metrics, who regularly makes speaking appearances.

Mike's previous appearances include Text Analytics World, Rutgers Business School, SEMRush webinar, ClickZ Live.

Mike also founded and writes for Biznology, is the co-author of Outside-In Marketing (with James Mathewson) and the best-selling Search Engine Marketing, Inc. (now in its 3rd edition, and sole author of Do It Wrong Quickly, named by the Miami Herald as one of the 11 best business books of 2007.


Great perspective, Mike. You say... "But what if you know all this already..." and then fill in the blank w/ Web metrics. Ok but I'm confident you'll agree -- defining conversions is a good "first step" but in urgent times like these marketers are challenged to go further. Specifically, the obvious next step is to organize around behavioral conversions -- whether you're a retailer trying to induce a sale or looking to generate leads or trials that ultimately convert into a customer. This involves connecting Web metrics to business metrics and creating systems across the enterprise (not just marketing!!) to drive behavior. Marketing (the discipline) becomes shared.

Leading retail and B2B marketers are understanding the "chronology of purchase intent" and organizing around it to drive actions not just create "branding" through reach-and-frequency. Think direct marketing right? Then apply it everywhere.

As marketers we'll be paid less (in 2009 and beyond) for our ability to move customers toward a purchase and more based on the ACTIONS we create -- and measure. It all must be measured as you illustrate perfectly. Marketing accountability but across the enterprise!

More evidence from 2008 Customer Experience Impact Report (Harris Interactive):
Business leaders expect contact centers "to move from being a cost center that answers questions and handles complaints, to a more dynamic part of the organization—supporting the customer and generating revenue."

I didn't know we could measure our conversion ratio offline, how would I do it?

P.S. Analytics is great, I'm using it, although I'm not sure if it's even 80% accurate... what do you say?

Hi Dominic,

You measure your offline conversions using your standard offline sales metrics. It depends on the business how you do it, but most businesses can measure leads passed to the sales force, or phone calls to the sales center, or some other first step offline that they can later tie to an offline sale.

One of the major problems with tracking online conversions is how accurate they really are. I currently have two specific products that were converting really well and all the sudden are not converting at all. It is a bit strange.

I couldn't imagine not measuring conversions at CyberCucina; it would be like driving blind!

There is no excuse for a company not to have a basic web analytics platform that provides insight into site traffic patterns, referring sites, and search phrases, especially since Google offers free Web analytics and free conversion tracking via AdWords.

Comments closed after 30 days to combat spam.

Search Engine Guide > Mike Moran > Is there a standard way to measure conversions?