Case Study: Calculating the Viral Rate

As we noted in our piece on "Big Seed Marketing," a marketing "virus" must infect one or more people. Anthing less than a 100% infection will peter out and eventually die off. As Watts and Perreti note, designing a truly "viral" campaign is almost impossible. Most viral efforts are flukes and exceptions. Better, say the authors, to hit a large swath of people with a message, hope for a decent viral rate and then gather in more people as the message slowly decays back to parity.

Here is an update on a case study I published last week (a petition about an Oprah show), the results we saw from our petition, and how we calculated the "viral" or reproduction rate.

By any measure, the petition has been a success. To date 21,000+ people have signed on. This from a seedbase of no more than 7000 emails. So, what was the viral rate of this campaign? It's more difficult to calculate than you think. First let's set up some definitions:

• General Viral Rate - This is calculated by taking the number of recipients divided by the number of people infected over one generation of the email. So, for example, if I send an email to 10 people and half of those people send it on to one other person then the viral rate is 50%. Without using software such as ForwardTrack this number will be an estimation because we don't know the number of generations that have been processed nor the true number of "opens" and forwards.
• Impression Viral Rate - This percentage narrows our focus to the number of people who have actually opened the email and sent it on to other people who actually opened the email.
• Click-through Viral Rate - Now we're getting down to those who actually clicked through to a landing page to take action.
• Conversion Viral Rate - Lastly, we're looking for the % of people who actually signed on to the petition.

This is a fairly young science and we're treading on new ground here. Some of these rates we can calculate pretty accurately. Others, not so much.

Here are the facts on the ground frame from Phase 1 of our (which ran from 10/3 through 10/9):

Recipients: 6871

Total opens: 4136

total click-throughs: 1912

Unique petition page visitors: 25000

Conversion rate: 45%

Sign-ups: 11000

Here's a verbal walkthrough: On October 3rd we sent out our petition request to 6800 recipients. Over the next 4 days the original emails were opened by 4136 people (this includes the original recipients and people that received the email in return) and nearly 2000 people clicked through the links in those emails. In the end, however, 25,000 people visited the petition page and 11,000 of those people signed the petition (a 45% conversion rate). Apparently, people created their own emails and forwarded the link onto their friends. (Note: we use iContact for our email system and Google Analytics on the website).

There are quite a few unknowns here. So let's go with the three facts that we know for sure: number of original recipients, unique visitors to the pettion form and petition sign-ups over those 4 days.

RECIPIENTS OR OVERALL VIRAL RATE

Basically, we want to determine what the infection or viral rate would have to be on the original number of recipients (6800) to get to 25,000 visitors on the petition page. There are two ways to calculate this. 1) use the 6871 number as assumptive starting point or 2) assume a 30% open rate and bring the starting number down to 2061. In other words: should the viral rate be calculated to the overall bucket of recipients or the assumed 30% open rate? Well, let's do both.

Total Recipients To get to the magic number of 25,000 unique petition page visitors we come up with a viral rate of 78.5% over 30 generations. Which looks something like this:

Basically, this chart shows a total accumulation of 25,000 people by the 30th generation. Of course this assumes that everyone clicked through to the petition who received the email. This is the extreme case and unlikely but this model gives you an idea of what were doing here.

30% Opens

Starting at 2000 who opened the email we need an average viral rate of 91.9% over 50 generations to get to 25K.

Of course, we don't know what the viral rate was between successive generations and it's likely that the open rate was different for different people.

Alternate Calculations

One alternate theory is that the first generation email of Phase I was above 100% on average and that the rate slowly degraded from there. This is possible but again difficult to calculate.

Lastly, we can work ourselves backwards by increments and guestimate what happened:

1. 11,000 people signed the petition during Phase 1 (solid number)
2. 25,000 people visited the petition page with 45% of these converting (signing ) the petition (solid number)
3. Assuming a 25% click-through rate we get 125,000 emails that were opened (estimation)
4. Starting with the 6800 recipients we would need average viral/reproduction rate of 94.5% over 100 generations to get to the 125K number.

My conclusion is that our viral rate was probably above 80% and below 100%.

Next Steps

In Phase II of our campaign we sent a follow-up email to the 11,000 signs-ups and encouraged them to forward it on to two other people. After another 4 days we had essentially doubled the petition sign-ups. Working off the same model of total recipients and 30% open rate we came up with 51% viral rate over 10 generations and 84% over 45 generations respectively. As you can guess ,the open rates and click throughs were lower for phase II as would be expected.

Rinse, lather, repeat. Tomorrow (Monday) we will send another email out to the new 11,000 emails who signed up over Phase II and try to anticipate the number of new sign-ups. My guess is that we will see similar results but a decaying number of sign-ups. We've put a few new measures in place to track things more accurately. Stay tuned.