Phone reservations are difficult and not worth it.

I have had a number of people reserve by phone. Every one of those people has cancelled or simply failed to follow up and de facto cancelled without alerting me.
My hypothesis is that the web site is clear and understandable, the product is simple and well-explained.  People who see the product and decide they want a bike just reserve using the online system / email. Done. prob (purchase) = 1.
People who see the product and are unsure have to call to better understand what we're doing, or to help assuage the uncertainty. Perhaps they need a harder "sell"? In any event, they have a prob(purchase) < 1, so far prob (purchase) = 0, with a small sample size. So I'll say they have 0 < prob (purchase) < 1, with sample size of 4 (I think the central limit theorem would fail here, so hard to estimate the true confidence).
These considerations lead me to believe that phone reservations are unlikely to result in a sale, and they are more costly in terms to time, so the expected value of phone reservations is very low. Lets say the prob (purchase) = 0.5 and the cost of the phone work is $20.
Then EV = 0.5 * 80 - 20 = $20.
Just not worth it.
On the other hand an online sale requires so work up front and has prob(purchase) = 0.9 after the reservation is booked.
EV = 0.9 * 80 = $72.
That's why I've discouraged them in the future, placing the phone number in a somewhat difficult to find place and only allowing email or online reservations.