Introduction
To produce AI insights into what your customers think, we need to capture 20 responses (which are at least three words long) from your customers. We do this by asking a small percentage of your returning customers, on a scale of 0 - 10, how likely they are to refer a friend. Once they leave a score, we ask them to tell us why they gave that score.
Customers who see a survey before seeing a referral offer are slightly less likely to refer than those who only saw a referral offer. When choosing what percentage of customers should see the survey first, you need to strike a balance between populating your dashboard quickly versus affecting your referral scheme. The higher the percentage, the quicker time to seeing results, but the bigger the impact on your referral scheme each month.
Estimating typical monthly orders and who sees a survey post purchase
Using past data, we predict how many customers you can expect in a month and how many of these are customers who’ve ordered before. We take the percentage split of these first time customers to work out how many customers per month are likely to see a survey. This split is selected during survey set up.
Impact on new customers acquired through referral
Customers are asked how likely they are to refer, on a scale of 0 to 10, with 10 being the most likely. We generally suggest that customers who leave a low NPS score are not shown a referral offer. Seeing a referral offer after leaving a low score can make the customer feel like they’re not being listened to.
During set up, you have the option to set the threshold that sees referral to either a score of 7 or above or 9 or above. We recommend setting the threshold to 7 or above to maximise the number of customers that see referral.
For customers who do see a referral offer, they are still less likely to refer. From previous data, we know the referral conversion rate for customers who have seen a survey is slightly lower. It’s about 80% of your typical conversion rate. We can estimate this small change in conversion unique to those customers who see a survey. This means the overall impact this has on customer acquisition can be calculated.
Time until calculating your untracked referral
We need 20 responses of sufficient length to produce AI insights. Of customers who see a survey, we use past data from other merchants to work out how likely they are to answer. The time to results prediction is based on your estimated monthly customers seeing a survey multiplied by likelihood to answer.
