From my previous visits, the meetups always felt to have a fun mix of people with different experience levels in crowdsourcing. To address this, I tried to keep the talk general touching on points I thought would be interesting to everyone.
I believe the tactic worked well because after the talk I was approached by a number of people who wanted to speak with me more about the topic, get advice on projects they’re evaluating and discuss possible partnerships. Basically, people who think crowdsourcing is as awesome as I do.
The event was recorded, so I will post the video when available. In the meantime, here’s gist of what I aimed to discuss:
Topic: Dispelling Disbelief and Building Trust in CrowdSourcing
I firmly believe that crowdsourcing as a business methodology is early in the adoption cycle and so we as crowdsourcing companies need to be sure we position our business approach correctly.
To help provide a bit more context, this video clip from the show Archer reflects the common prevailing perception of crowdsourcing:
A Common View About Crowdsourcing
As you can see, there’s a [hilarious] guy in a suit just standing before a group prompting them for ideas. There isn’t much structure, it’s chaotic and ultimately nothing of value comes of it.
Thankfully, businesses using crowdsourcing are starting to frame analogies that better represent the reality of crowdsourcing. One common analogy is ‘The Factory.’
The Factory is a very interesting analogy because it illustrates that crowdsourcing [done correctly]:
- Assembly Lines
- Constant Quality Checking
- Specialization of Workers
In my opinion, there are some serious problems with this analogy.
Crowdsourcing already has a very bad social stigma of equating to low pay (if any) and exploitative business practices. These are the same issues that people have with factories too.
The high level bad baggage that factories connote are:
- Sweat Shops
- Low Pay
- Low Skill
- Shoddy Work
- Outsourcing (and to some extent the decline of Western Power)
So how can we distance ourselves from these social stigmas?
The answer is to use a better analogy and one that really reflects the value and potential of crowdsourcing [done correctly].
Queue solution: “The Orchestra”
- Skilled Participants
- Focused Effort
- High Quality
This means that the musicians are the crowd. But what does that make the crowdsourcing company?
- Constant Quality Checking
- Specialization of Workers
I’ve personally used this analogy a number of times while talking with prospective business partners and it’s been extremely effective at achieving the same “ah, I get it now” reaction as the factory but it simultaneously distances crowdsourcing from the concept of ‘low pay, low skill, shitty quality.’
When building trust with businesses, it’s fine to have the high-level discussion but we invariably have to turn to specifics. And to help with that, here were three easy to remember points that I believe are absolutely critical every good crowdsourcing company (including contest and game companies) need to have in place:
1. Mimic the real world by creating social accountability:
- Disallow future access to worker. This is probably the most basic form of social accountability.
- Reputation Systems: These are becoming more and more common in the crowdsourcing space because they’re extremely powerful for maintaining “crowd accountability.” A reputation system can be as simple as something we see on Ebay with number of sales, number of positive reviews or something in-depth like on Elance, Guru and Trada which all maintain a slew of metrics to help contextualize the social standing of a crowdsourced labor provider (i.e. workers). A reputation system can also be a blend, using statistics to gauge worker reliability. Whichever route you go, be sure to maintain social accountability.
2. Be sure that quality control is part of the design process first, not an after thought
- Thinking back in terms of the orchestra, we need to be sure play the role of the conductor and maintain control over the quality. As the crowdsourcing company (or participant if we’re simply hosting a contest or creating a game), if we forget to account for methods of controlling quality, we’ll quickly run into trouble due to the magnitude (i.e. scale) of work produced by the crowd.
- A common quality control method is peer review. This is seen at most companies ranging from Clickworker to Servio and is extremely effective.
- Another equally effective method of controlling quality is by gating users. Simple examples include ‘pre-tests’ like you’ll see with Crowdflower tasks on Amazon Mechanical Turk or credential tests as used by Servio. Other curation techniques exist and there is discussion whether this is ‘real crowdsourcing.’ It’s more a question of ‘pure’ versus ‘practical’ in my opinion.
- Ultimately, you need to have some form of quality control up front and make sure it’s not an afterthought.
3. Don’t rush in blindly
- It’s easy to think that crowdsourcing is easy: “All I have to do is put the work up and people will do it. Done.”
- The reality of crowdsourcing is that it is not straight forward. Maintaining and building crowds takes time, money and commitment. Do not underestimate the commitment. This means, essentially, make sure you have a real business model in place that will sustain the crowd. If all you do is get one deal and no prospects of a future deal using the same skill set your developing in your crowd, then you’ll likely be wasting time and money that could be better spent in other areas. I’ve been down this road in the past and it’s very painful. The benefits of crowdsourcing are scale and consistency. If you don’t have a consistent flow of work to “feed the crowd,” then you’ll never achieve the level of scale and consistency that crowdsourcing can provide. You’ll basically just be a faking it–which means that you’ll really be following old school business processes (using freelancers or building internal teams) but claiming to be a crowdsourcing company. So, be sure to think about whether you’re truly committed to the business model and the crowd before jumping in.