ellipsis flag icon-blogicon-check icon-comments icon-email icon-error icon-facebook icon-follow-comment icon-googleicon-hamburger icon-imedia-blog icon-imediaicon-instagramicon-left-arrow icon-linked-in icon-linked icon-linkedin icon-multi-page-view icon-person icon-print icon-right-arrow icon-save icon-searchicon-share-arrow icon-single-page-view icon-tag icon-twitter icon-unfollow icon-upload icon-valid icon-video-play icon-views icon-website icon-youtubelogo-imedia-white logo-imedia logo-mediaWhite review-star thumbs_down thumbs_up

Yes, You Can Predict Viral Marketing

Yes, You Can Predict Viral Marketing Joseph Carrabis

A reader recently contacted me about my columns, asking how to get a handle on what to expect and when from a viral campaign. What follows is based on that conversation and the NextStage research that inspired those two columns.


It is basic and should provide an outline to consider before getting involved in a marketing campaign, viral or otherwise.


What's completely controllable and what's not
The completely controllable part of viral marketing is based on mathematics and deals with the following questions:



  1. How many individuals does the campaign need to start with (seed)?

  2. How fast will the campaign spread (propagation factor)?

  3. How will the campaign spread (vectors)?

  4. How large a group is required to sustain the propagation (viral burden)?

  5. What is the campaign's goal (maintenance factor)?

  6. How large a group is required to sustain the campaign once the goal is achieved (threshold point)?

  7. At what point is the campaign too successful (saturation point)?

There are many more questions than what's listed above, and (again) we're showing a simplified process.


What's less controllable deals with Jim Meskauskas' .


Many people feel this is not controllable. I disagree. Entertainment through uniqueness, trust and fair-exchange are both predictable and probabilistic using the right tools to determine audience-to-message match and related factors. I'm going to call this the Meskauskas-Carrabis Effect, or MCE.


Foresight versus hindsight
Before going further I'm going to introduce two factors, one from psychology called
Hindsight Bias and one from economics called The Endowment Effect. These two factors play heavily when predicting the outcome of any efforts, marketing or otherwise.


Hindsight Bias occurs when, given an equal probability of several different outcomes, a single outcome occurs and people say, "I knew that was going to happen." The truth is they didn't know per se, only that once a single outcome occurs they can easily reconstruct the path that led to that singular outcome, hence it seems obvious.


The Endowment Effect can be summed up as "The devil you know is better than the devil you don't" crossed with "A bird in the hand is worth two in the bush." It comes into play when people are either undergoing or planning change, and it causes people to believe that what they have is better than what they will have or will get.


Hindsight Bias and the Endowment Effect can make some startlingly accurate predictions about viral marketing outcomes. Here's how it's done.


Start at the end and work back to the beginning
Start by asking, "What is the end result of this campaign?" Alternately, you can ask, "How will we know when this campaign is successful?" Often people will answer nebulously with something like "Sales will double" or "We get a million downloads," neither of which is good enough.


When will sales double? Let's say in six months. Now we need to know the current conversion rate. We'll use round numbers and say 20 percent. Based on nothing other than traffic volume, it would need to increase by 100 percent for sales to double. But this can be misleading. The conversion rate is still 20 percent, even though sales have doubled, so the question remains, "How does traffic increase by 100 percent?"


Precision descriptions
Precisely describe the end result of the campaign. Doing so ensures being able to recognize when the end is reached -- thus avoiding scope creep -- or that the desired end result is going further and further away as the campaign progresses, therefore it's time to re-evaluate the goals.


Now that the end result is documented, turn things around. The exactingly described end result becomes the starting point; the existing situation becomes the end point.


Precisely describe the existing situation. This is where most companies fall down. They know where they are but can't describe their exact position or what got them there with any degree of accuracy.


Ask, "How did this success occur?"
Imagine the campaign being the desired success and ask: "What had to happen right before we realized we were successful?"


As before, precisely define what needs to happen. Once that's done, do it again. This time ask, "What had to happen right before we realized we were on the verge of success?"


Repeat this "stepping back" process until the campaign is completely defined from end result to current market situation. This process defines the campaign's milestones and benchmarks by



  1. Using Hindsight Bias to predict what needs to happen before it needs to happen

  2. Documenting which tools will be required and when

  3. Providing alternative tactics and strategies to get the campaign back on course when strangeness occurs, and

  4. Avoiding the Endowment Effect by working backwards from what will happen to what needs to happen for goals to be reached.

These suggestions apply to any campaign and comprise the following simple steps:



  1. Clearly define the end goal. Write it down as a paragraph and draw a diagram (sales chart, geographic or demographic penetration, et cetera) that demonstrates the goal as recognizable numbers.

  2. Clearly define the current situation. Write it down as a paragraph and draw a similar diagram that demonstrates the current state as recognizable numbers.

  3. Begin "stepping back" from the goal. Use as many steps as necessary and ignore any time-management requirements at present. The purpose is to clearly "look back" from the goal to the current state and document each milestone or benchmark that existed between the goal and current state, why each was important, how they occurred, how they were passed or exceeded and what was required to ensure each step's success.


Let's say the end goal is to have a self-sustaining viral campaign in which 100,000 unique visitors download a particular tool each month. Those 100,000 visitors are the ones acting upon the viral suggestion. The step back from that is, "How many people are necessary to sustain that conversion rate?" One hundred thousand visitors represent 20 percent of the traffic using the numbers given above. This means 500,000 -- half a million -- unique visitors per month are required to sustain 100,000 downloads a month.


But we're dealing with a viral campaign, so sustaining the conversion rate is actually secondary to, "How many people need to receive the message in order for half a million people to visit the site each month?"


This question is the next step back, and it is where the seven questions at the start of this column start getting answered.


Now determine how the viral message is to be transmitted. Word of Mouth alone? Will there be an email component? Will some other transmission vector be involved?


At this point some of the NextStage research I've discussed previously can come into play: does the transmission method involve

This means the viral message needs to start in (for example) pure word of mouth, then change to email and word of mouth, and then change to pure email at some point in time between the start of the campaign and its fifth iteration.


If you send an email to a friend using the rules listed here, the campaign needs to switch vectors right after your friend's friend's friend's friend tells their friend in order for the campaign to continue to propagate.


The next step back is, "How fast should the message spread?" This is a critical question because there is such a thing as too fast. An excellent message that spreads too quickly gets lost because it doesn't stay in anyone's mind long enough to make an impression. Too slow and the best message never gets reinforced enough to be acted upon. Using our above numbers, decisions about how soon from the start of the campaign to 100,000 downloads a month needs to happen and how long it should continue need to be made.


The last question that needs to be answered in this simplified schema takes us back to the beginning of the campaign: How many individuals are required to start the campaign in order to meet the campaign's goals?


Our example above started with a single person and assumed each person receiving the message sent it on to a uniform 10 contacts. We didn't take into account people who are "immune" to the campaign or who pass the message on but do not act upon it themselves.


Starting the campaign with one person and using the 20 percent conversion factor means the message never gets out or does so too slowly to be useful. Time-to-fruition and end-goal requirements help pinpoint what the seed population needs to be and some good rules of thumb are:


(Large end population) x (long time to fruition) = 12 to 15 percent of end population as seed
(Large end population) x (short time to fruition) = 25 percent of end population as seed


The catalyst in these formulae is the MCE. A high MCE equates to a shorter time to fruition or smaller seed. A low MCE means longer time to fruition or larger seed. Organic growth of a viral message can work very well, but only if you're providing something an audience -- which might not know you exist -- wants and is willing to go looking for it, and this equates to a very high MCE.


Note: If you're in Boston on August 17, 2006, I'll be presenting research on "Increasing Knowledge Transfer by Adapting Information Presentation Style on the Fly" at the Boston KM Forum.


Joseph Carrabis is CRO and Founder of NextStage Evolution and NextStage Global. Read full bio.

Joseph Carrabis is Founder and CRO of The NextStage Companies, NextStage Global and NextStage Analytics, companies that specialize in helping clients improve their marketing efforts and understand customer behavior. He's also applied neuroscience,...

View full biography

Comments

to leave comments.