Experiments With Open Participation
3.5
Friction is often considered undesirable. This section devotes a whole chapter to the removal of barriers in order to encourage repeatable inter- actions. However, friction can also serve as a source of value by discouraging the repeatability of undesirable interactions.
Traction and friction are at odds with each other. Friction, arising from intentional and unintentional barriers to participation on the platform, gets in the way of traction – large-scale participation – on the platform.
Friction may be created by design (for instance, producers may have to go through background checks before they get access to the platform) or by accident (for example, the interface may have poor navigability or usability).
Well-designed friction can lead to healthy and desirable interactions on the platform. Accidental and unintentional friction only gets in the way of trac- tion, without creating additional value. Getting friction right is critical to the success of any plug-and-play architecture. This chapter explores some of the key design considerations when designing friction into the platform.
Platforms must be designed in a manner that optimally balances the quality
and quantity of interactions by balancing traction and friction.
As with all design considerations, the ultimate goal of a platform is to enable repeatable interactions. Hence, as a rule of thumb:
Friction is a good thing only if it increases the repeatability of desirable interactions and decreases the repeatability of undesirable ones.
the traction-friction matrix
The trade-off between traction and friction is best visualized through the traction-friction matrix.
• High friction–low traction: A platform may lie in this quadrant for one of two reasons: by design or by accident. Platforms that need high curation in their initial days fall into this quadrant. Alternately, poor interface design and navigability may throw a platform into this quadrant.
• Low friction–high traction: A platform falls into this quadrant for one of two reasons: frictionless experiences by design or a general lack of checks and balances.
• High friction–high traction: This is a great place to be, and successful platforms eventually migrate to this quadrant after starting off in one of the two quadrants described above.
• Low friction–low traction: This is the worst quadrant to get stuck in for too long. Platforms experiencing the chicken-and-egg problem fall into this quadrant.
movements in the matrix
The matrix helps to lay out the various strategic choices that a platform may take to balance the contrasting goals of open participation and cura- tion that were explored in Section 2. This matrix provides a framework for understanding how different platforms achieve these divergent and conflicting goals.
1. Pivoting around friction
A platform may constantly need to juggle the quality of interactions with the number of interactions. Too much friction may lead to low participation
and fewer interactions. Too little friction may lead to a poor-quality culture on the platform and discourage quality producers in the long run. The platform must ensure that it achieves the right balance between friction and traction at every point in its journey to platform scale.
2. Avoiding friction altogether
Some platforms may avoid friction entirely. This is especially true for plat- forms with low interaction risks. For instance, Instagram’s adoption was driven by low friction in usage.
In contrast, platforms with high interaction risks may not benefit from avoiding friction. Craigslist enjoys strong network effects because it extends across categories with low as well as high interaction risks. The platform allows anyone to post a listing, without appropriate checks and balances.
While the lack of friction may work very well for low-risk categories and encourage interactions there, it may lead to undesirable and dangerous interactions in high-risk categories.
3. Embracing friction with scale
Quora has been increasing participation friction as it scales in a way that encourages quality on the platform. Anyone could ask a question in its early days, but asking a question on the platform now requires the user to pay in Quora credits. Promoting a question or answer on Quora may also be accomplished using Quora credits. As the platform has gained further traction, it has gradually embraced a higher level of friction. Content promotion increasingly requires more credits than it once did.
4. Relaxation of norms
App.net started with high friction, instituted through a $50 subscription fee. However, it has gradually reduced friction to allow for traction and eventually moved away from its original goal of offering an alternative to Twitter. In contrast, other platforms, like Sittercity, have succeeded despite high friction and never had to relax their norms.
5. Scaling the country club
Some invite-only platforms scale well in spite of high-friction. They start small but preserve high quality as they scale. Participants value quality, and the higher friction encourages the repeat participation of a curated set of participants while discouraging that of others.
design considerations for friction
Two platforms in the same vertical and category may compete and co-exist by staying in two different boxes in the traction-friction matrix, as the examples below demonstrate.
1. Friction as a source of quality
Undesirable interactions on a platform hinder the repeatability of desirable interactions. Women tend to avoid dating platforms that attract stalkers.
As the proportion of undesirable interactions increases, the repeatability of desirable interactions decreases further as high-quality producers and consumers rapidly abandon the platform.
Choosing a babysitter also benefits from high quality. False positives may lead to risky interactions. Friction, in the form of background checks on babysitters, provides a valuable source of information on quality. In contrast, the lack of friction on Craigslist makes it unsuitable for hiring babysitters online.
2. Friction as a source of superior signaling
Friction may also lead to better signaling. Background checks on babysit- ters yield exact parameters for parents to make a decision. Hence, the friction of curation helps with market signaling.
Employment markets also rely on signaling. The process of profile creation on LinkedIn is onerous and involves high friction. However, it helps the platform gather the requisite data required to signal candidate quality to recruiters. By breaking a high-friction and high-investment task into manageable chunks, LinkedIn successfully gathers rich data on every individual professional.
3. Friction as a barrier
For all the hype and fanfare surrounding App.net’s launch, the platform never quite lived up to its initial intent of providing an alternative to Twitter.
Two design considerations were flawed to begin with:
a) Applying friction to both producer and consumer roles: By requiring users to pay to access the platform, App.net added friction to both produc- tion and consumption. Twitter allows its producers to build a following.
By restricting overall access to App.net, the platform limited its ability to enable producers to gain a strong following.
b) Friction without a guarantee of quality or signal: While an access fee introduces friction on App.net, that friction doesn’t serve as a guarantee of quality or as a signaling mechanism. It doesn’t help to improve interac- tions on the platform. App.net realized that friction wasn’t helping and subsequently lowered the access fee, through a series of revisions, by more than 90%.
platform scale imperative
In the relentless quest for platform scale, one may often see frictionless interfaces as a necessary design principle. However, friction may serve to encourage the repeatability of desirable interactions by encouraging high- quality interactions. The following is a non-exhaustive list of design ques- tions to consider while introducing friction onto a platform.
a. Does the friction affect one side or both? Does friction on either side add value to the other side? Does friction on either side deplete value for the other side?
b. What are the sources of friction? Do they improve quality and help with signaling?
c. Does friction increase the repeatability of desirable interactions?
d. Is the interaction high-value or high-risk?