Thanxgiving: Viral Marketing Campaign Within Your Loyalty Program
Recently we launched our “Thanxgiving” feature, which is basically a viral marketing campaign where the inviter and the invitee get rewarded for sharing it. While this is pretty standard in the world of apps, it’s not exactly something you’re likely to encounter in a loyalty program at your favorite pizza shop.
Loyalty solutions are principally about retaining existing customers, but with Thanxgiving customers aren’t just inviting friends to sign up for the merchant’s loyalty program. They’re inviting them to come make a purchase and experience one of their favorite brands.
Key to how our program operates is that both parties get rewarded, but only after the invited person signs up and makes and purchase at the merchant. No one gets rewarded until the merchant sees actual revenue from the invited consumer.
The potential value here is huge, as research shows that consumers acquired this way are significantly more valuable than customers acquired elsewhere. They stick around 37% longer and spend roughly 150% more. So, how do we pick the perfect incentive to capture this extremely valuable demographic? That’s where data science comes in.
At Thanx, We Don’t Like Guessing – We Like Data
In designing Thanxgiving, we did a lot of research into similar viral marketing campaign programs. We couldn’t find anything that resembled the thing we wanted to build. While it’s exciting to be pioneers, this also presented something of a problem, as we had little information as to how incentives should be structured. Should we offer the person sending invitations 10% progress towards a reward? 50%? A full reward – a free pizza or a car wash for example? These incentives cost our merchant customers money and we don’t take that lightly. While Thanxgiving guarantees positive ROI because of the program structure, we don’t want to give unnecessary discounts. So, the question became this: What’s the optimal incentivize to get consumers to complete a referral?
Network Participation Is Good For Everyone
Instead of just guessing what the incentive should be, or asking each merchant to guess, we wanted to get hard data. We built the system to allow our merchants to choose both rewards – what the inviter gets for signing up a friend and what the friend gets. So for the first 6 weeks after we launched the feature across all our merchants, we tested it. This is one of the many values that merchants get by sharing the same platform. We can get data to answer questions like these and improve outcomes for ALL Thanx merchants’ loyalty incentive programs (and viral marketing campaigns).
Each consumer on the Thanx platform was divided into one of 4 buckets to determine what incentive they’d get for signing up a friend:
- 10% progress towards a reward
- 25% progress towards a reward
- 50% progress towards a reward
- 100% progress (a whole reward – a free sandwich or a $10 discount for example)
Unsurprisingly, the more generous the incentive, the more likely a customer was to send an invite. Customers who were offered 100% progress (a whole reward) were 60% more likely to send an invitation. Interestingly, the difference in participation between 50% progress and a whole reward was nearly non-existent:
% of consumers who sent at least one invitation
The incentive given to the inviter had a clear impact on how many people they were able to enroll successfully. The friends these customers invited all got the same thing. We only tested the incentive given to consumers for inviting someone, not the incentive for signing up (don’t worry, we will).
Our big takeaway: the likelihood of a consumer to sign up was directly correlated to the value of the incentive given to the inviter. Inviters who were offered higher-value rewards were better at getting their friends to sign up.
Effectiveness of each group at signing up friends
Perhaps equally as telling is how likely it was that referrals would not just create an account, but actually come in and spend money. Spoiler: the bigger the incentive given to the inviter, the more likely their friends were to actually become customers. While conceptually this seems unlikely, the results are fairly clear:
Thanxgiving signups who made a purchase per 1000 active customers in each cohort
The bottom line here is that the difference in effectiveness between the 50% incentive and the 100% incentive is negligible. The difference between 10% and 50% though is fairly notable, the latter being roughly 80% more effective than the former. This is why Thanxgiving works: we know exactly what it takes to incentivize action, and we don’t waste our merchants’ resources throwing rewards at customers who don’t convert.
Data-Driven Results That Generate Actual Revenue
We resolved our test after signing up many thousands of consumers. All of our merchant’s referral programs now default to the 50% offering, though this is configurable by them.
This is perhaps the best part of working on the Thanx product. I’ve been building this product since before it launched and the data always manages to surprise me in some way. I’ve learned to be distrustful of my hunches and invest energy into experimentation and analysis. It’s rewarding when the result of that analysis translates into answers for customers and even more so when it translates into revenue for them.
Thanxgiving has generated thousands of dollars for all of our merchant partners with viral marketing campaigns… without them having to do anything. How cool is that?