Monday, November 28, 2011

Most likely your Conversion Rate is Wrong

I am sorry to break this news to you but chances are that your conversion numbers are all messed up. Let me demonstrate it with an example.

Let’s take a simple conversion process, see the figure below. Let’s assume that you get 100,000 visitors on your site, and 20,000 start the checkout process, 15,000 make it the next step and finally 5,000 end up converting.

Based on your web analytics tool your conversion rate will be 5,000/100,000 = 5% . So far everything looks good.





But what if your checkout path looks something like this?



Do you see the phone number? I bet many customers are using that phone number to complete the purchase. Are you taking those, who call and convert, into consideration when calculating the conversion rate? Do you know that when someone picks up the phone to call he/she has a higher likelihood of converting?

Taking the example above, here is how your conversion funnel looks like:


Most of the web analytics tools just allow you to see a view of single channel conversion rate i.e. web conversion rate. However, as I discussed in my post "Are you Optimizing the Wrong Steps of the Conversion Process?", customers don’t care how your channels are divided or who is responsible for what channel at your organization. They care about their money and will use whatever channel they feel most comfortable with.

Are you going beyond single channel when calculating your conversion rate? If your answer is no then your conversion rate is wrong.

Thoughts? Comments?

------------------------------------------------------------------------------------------------------------------

Web Analytics Jobs

8 comments:

  1. My thought: only include the people that have a 'buy' intention (people who look at a product or people who put something in their basket). By including all visitors that hit the site you also include people who only want your address. It's not fair to include them in conversion rate, and that breaks conversion rate also.

    ReplyDelete
  2. Anonymous9:13 AM

    Interesting point, but I might suggest that the phone-number "path" really depends on the type of site. For instance, I'd guess that consumers are somewhat trained to realize that using a phone-number for an online-only store is not going to be an easy process, but using a phone-number for a B&M store would be easier. Obvioulsy, this is a general, personal observation, but it seems like one would have to take these site differences into consideration, too, when looking at conversion rates. I also wonder if research has been done on the use of phone numbers (from websites) to complete orders.

    ReplyDelete
  3. Andrea, I agree that there are people who are not in the market or have already converted. Maybe you want to exclude them from the calculation. I wrote a blog post earlier on similar topic, http://webanalysis.blogspot.com/2010/01/is-your-conversion-rate-wrong.html#axzz1f1bCWDa6

    You should choose your denominator (total site visits or visitors excluding those who have no intention of converting etc.) considering your site, organization and make sure everybody in the organization is on the same page. Most widely used calculations just use total site visits or visitors though.

    ReplyDelete
  4. @Ananymous Yes how many conversions you get via phone or how many people end up using the phone number will depend on the site type, how visible the phone number is, market perception about how easy the phone call will be etc.

    We have looked at some numbers and found that in one case 25% of the sales were completed on Phone and another 25% were assisted by phone. I will see if I can find some publicly available research.

    ReplyDelete
  5. Don't forget about those people who do their research online and drive to the store to make the purchase. That could add another couple thousand conversions, so now you have 5,000 online conversions + 2,000 phone conversions + 2,000 in-store conversions = 9,000 total conversions, or 9%.

    Obviously this depends on the type of site and industry, but I can see how important this information would be to national retailers (i.e. Walmart, Best Buy, etc.). This raises an interesting point: are we cannibalizing our phone and in-store conversions when we increase our online conversion rate? And if so, is a visitor who converts online more valuable than someone who calls or visits the brick and morter store?

    On one hand, with the online conversion, we gain the ability to market to them after the sale (i.e. email marketing). On the other hand, if they visit the store in person, I would *assume* there's a higher chance that they'll purchase more - either impulse buys or for a number of other reasons.

    Food for thought :)

    ReplyDelete
  6. @anil Our mobile sites now record emails, ringbacks (where we pay for the call) and tap/click to call phone volumes.

    It's AWESOME. About 2-7x phone volumes and a conversion rate that's about 30%, not 8%.

    Your point is well made - conversion to phone is NOT the same as online transactions so there is a big variability in conversion rate - that you can't see if you don't track phone.

    This is my mission in 2011 - to get more people tracking phone and other contact/cost channels.

    C.

    ReplyDelete
  7. Anonymous9:13 AM

    @Jim. That's the key part, once you can attribute the journeys through the multichannel mix then you can measure lifetime value via each channel and optimise accordingly. Online is certainly easy to service cheaply however your segmentation may tell you that a certain group crave personal contact... Loyalty may be reduced.

    ReplyDelete
  8. We've been utilizing OpenVBX and Twilio to provision 800 numbers for our clients. The numbers are cheap and available on demand as new campaigns come up.

    We're still working on closing the last step in the sales funnel by tying the phoned from # to a final sale.

    Anyone have any other methods that they like for online to offline tracking?

    ReplyDelete

I would like to hear your comments and questions.