If you’re a data geek like me, then this one is for you. Let’s say you have a business and you’re trying to market a product or service. You know your marketing is working since people are responding at a fairly predictable rate, but you can’t figure out why your sales are all over the place. You, my friend, are seeing the gap between response and conversion.

There is often a disconnect between response (i.e., people that respond to your marketing) and conversion (i.e., of the people that respond, the ones who buy) that annoys marketers everywhere, yet there are proven methods to help make these two marketing metrics more equally predictable. The solution is lead nurturing and there are some things you need to know in order to effectively do it.

But first, let’s explore a bit further into why this gap in predictability of response and conversion exists.

WARNING: This is about to get geeky, but I’ll keep it digestible.

To keep it simple, let’s look at an example. Say you send out 200,000 emails and get a response rate of 0.5%. How likely is that to occur again? Well, without going into a full blown statistics lesson, we can use a special calculator to predict the results if we were going to run the identical marketing program again.

By entering the amount of emails you send and the response rate, the calculator will show you what the response rate range is at a 95% confidence level. So for our example of 0.5% response rate for 200,000 emails, we calculate that our response rate range is between .469% and .531% at a 95% confidence level. In other words, if we were able to travel back in time and run the exact same campaign to exactly the same people, the response rate would be between .469% and .531%, 95 out of 100 times.

Remember, we’re trying to figure out why there is a disconnect between response and conversion rate predictability. To answer that, let’s stick to our example and move down the sales funnel.

### Moving Down the Sales Funnel

Our range of response rates says that we’ll get between 938 and 1,062 leads. No riddles there. In fact, that’s a pretty narrow range that we can be happy with as marketers. The good thing is that we now have an pretty good idea of how many leads we’ll get if we run a similar program.

But here’s where things get funky. When we move down the funnel to predict the number of sales using the same calculator and same confidence interval of 95% a much less predictable range is discovered. First, which number do you use for your sample size in the calculator—938 or 1,062? Well, to be fair you have to use both numbers to figure out the high and low end of your possible sales.

So, let’s say the average conversion rate for your sales team is 30%. At a 95% confidence level and 938 leads you could have between 27.1% and 32.9% conversion rate. For the higher end of response at 1,062 it’s between 27.5% and 32.8% conversion rate.

Going one step further and calculating sales, 938 leads gives you between 254 and 303 sales. For 1,062 leads, you could have between 292 and 348 sales.

Look at the image below. Can you see how the number of sales becomes less and less predictable as you move down the sales funnel?

Now imagine if you were trying to predict your ROI. It’s impossible to accurately predict it because your range of sales is too great. Look at our example again and let’s say your CPM is \$500, giving you a total marketing cost of \$100,000. For this example assume your average revenue per sale is \$1,000. This means that your predicted revenue could be between \$254,000 and \$348,000. That’s a big gap and when you look at the total difference in predictable ROI it’s between 154% and 248%! As you can see, it’s really a useless way to estimate future performance.

### What This All Means

A lot of marketers try to use ROI as a way to select lists to mail or email to. In other words, they will rent or buy more records from a list that gives them historically better ROI than other ones. This is often a flawed way of thinking unless you have vast quantities to give you ranges of ROIs (like explained above) that are clearly higher or lower than each other. In most B2B situations, you simply won’t have enough quantity to definitively predict if one list will preform better than another in terms of ROI.

So what are you supposed to do? There are couple things sophisticated marketers are doing today in light of this information.

### Optimize on Response Rate

When you rent or purchase your lists, make your selections based on historical response rates rather than sales and ROI. This is a more reliable predictor of performance as I illustrated above. Then, move on to my next point.