Strategies for a referral system / viral growth: what worked for you?
I'm contemplating setting up a referral system for Monitoro, to leverage users' social networks .
There are many possible strategies and I'm unsure what approach is better (or the individual tradeoffs). I'm considering the two alternatives:
- Give credits redeemable through subscriptions (e.g. 1 month of free premium if you bring 5 users)
- Improve the current plan (e.g. 1 extra monitor for each new user)
Also wondering whether the referral should kick off at paid users, or any referred user. In the latter case I am expecting I'll need to improve my spam-detection game, so any tips in that direction would be appreciated if you think it's a better option.
Hustle your first 100-500 users first. Analyze their usage patterns and demographics. Implement a referral system or incentive plan once you see a significant result (i.e. Activeness)
Thank you @felix12777 for your insights!
What has been your approach so far to analyze usage patterns and demographic?
Automating Twitter using Jarvee has been very effective for me.
Thank you @Avatar for your answer! Could you share a bit more about your experience using twitter automation for referrals?
Sure, I've been using Jarvee to automate DM'ing. So I've set it up to target followers of my competitors and also relevant hashtags. Twitter is pretty generous with automation so it allows you to send a lot of messages each day.
Giving a referral bonus from non-paid users can bring A LOT of abuse. People can just use their 2nd email address or create a new every month, and ride free premium forever. And timewasters. People will ask their friends to sign-up just so that they get the bonus, even though they know there is a 0% chance their friends are going to really use or pay.
Also, a referral bonus that requires bringing 5 users may be too big "cliff." Until your referral system is proven, it may be easier to start with giving a bonus already from bringing 1 referral (especially if combined with the paid account required).
How are you planning to technically implement the referral thingy?
However my strategy so far for abuse management is to track all potential avenues of it (overusing the interactive editor for example), and take action when necessary.
The numbers look good so far, I'll implement proper countermeasures if this changes.
Generating a referral link with some ID for each user, and storing "referrerId" in the database when a new user signs up through that link (and a bit of logic to give the reward and track the event).
I use Metabase + SQL queries for analytics and to track user behaviour, and I can use it to identify anomalies and abusive referral chains.
But to be honest, it all sounds overkill at this stage. Abuse means usage/use case, and if I can understand the pattern behind it I might potentially find out a valuable niche.
So, I think we talk about the exact same thing, except the opposite :)
The "abuse" that I meant was "gaming" of the referral system to get the bonuses, by opening fake non-paid accounts (and not even from free email services) or asking friends to open accounts that they won't ever be converting to paid. So the behavior you'd need to detect is not-doing-anything , which is easy of course, but impossible to separate from legit not-doing-anything.
As background, one of my earlier startups got maybe 25% of growth via a referral system, and even though we gave the bonus only from a referral of paid sign-up, there were STILL people who tried to abuse the system.
And thanks for the tech details. Though I was hoping for a turn-key solution, I really don't want to spend the time coding referral and abuse detection systems when I should be focusing on the actual product.
It's really not that big of a deal. If someone is inviting too many users and ends up actually using the extra monitors, there's a high likelihood they'll upgrade to faster monitors (it's another dimension you cannot improve via referrals).
If they still don't, it would be interesting still to see what they are doing and potentially monetize that.
tl;dr, I'm not currently concerned about abuse.
I agree with you in general, and I started the product with as many turn key elements as possible, but soon enough I found myself needing control I couldn't get via external systems (and the bills build up incredibly quickly).
My current solution of choice is consolidating all user behaviour in a single place (metabase) and it's actually much cheaper to extend this setup than integrating a 3rd party system at this stage, while still keeping the exact flexibility I need.
For instance, I modeled the entire user journey, including a conversion funnel for the different user stages (that I define on the spot in the SQL query).
Imo having a good overview on what's going on is more essential than a featureful but ultimately less guided product, and is worth spending time for.
In my previous startup, we had a significant monthly $ cost per user. That probably explains our different concerns/priorities regarding abuse.
Regarding Metabase: hadn't heard about it before, seems interesting. What were your main reasons for choosing it over Amplitude or Mixpanel?
Here's my overall strategy from an Analytics perspective:
I have several types of events to track:
My event collection stack right now is pretty simple:
By consolidating these two sources, I am able to have a single overview, and:
I'm performing the actual analysis with two tools:
I'm using a variation of this approach since quite some time to perform business analysis in my day job, and I found it to be powerful transposed to this problem space.
Why am I not using Amplitude or Mixpanel or tools like that?
These tools give me mostly two things, a way to collect events and insightful queries (with their corresponding charts).
I'm satisfied with the current way I am collecting events. Regarding the insightful questions, I tried to build queries on Metabase that are relevant to the current challenges I am facing. It wouldn't hurt though to get inspiration from them and I am grateful you exposed me to these two.
Let me know if you'd be interested to read about this stack, I have considered writing about it (a variation of "How to analyze user activity for free").
Many thanks for the detailed info.
I'm doing the exact same: combining web and product analytics - and it's super important to do that.
We just implement it in the opposite ways: your analytics software works by pulling information from a database, while my analytics works by pushing events to the analytics software.
However, the big difference is that I cannot do free SQL queries or add new chart types to Amplitude. And that some soon-to-be necessary features like cohort analysis are not available in the free plan, but require subscription costing tens of thousands.
Regarding writing about your stack: yes, I'd definitely be interested in reading more about it, or alternatively maybe just buying your consulting time and get it done, if that is a service you offer?
Do you think this problem is abated if it requires sms verification instead of just email verification?
SMS might work well. I haven't done that though, so no real experience, only guess.
Just a "tell-a-friend" form
Great idea. I'll try it first before providing incentives.
Without a few thousand current users don't expect a referral program to move the needle. If this is important to you, build a lightweight version.
Thanks @gator70 for your insights!
The fact that you have a free plan makes this a bit harder. In general the better the offer, the more motivated users will be to refer others. I don't think you would see any success with 1 month of premium if you bring on 5 users, that's just a lot of friends for a tiny benefit. I'd go with your latter plan, that seems a lot more actionable, and motivating.
Thank you @shanefromfargo for your answer!
Your focus on it being actionable and motivating totally makes sense.
I am considering also suggesting to users to share some message to twitter with their referral URL. Do you think this social share alone should be rewarded already? Or only when their link brings referral users?
If you have little cost to host new users, I would make it as generous as possible. Then dial back as you iterate and find abuse.
Yeah this is completely aligned with my policy so far. Thanks for your confirmation.
This comment was deleted 6 years ago
Thank you @shiva_prabhakaran for your answer!
By transferrable, you mean keeping referral credits month-over-month right?
This comment was deleted 6 years ago