Sep 2, 2011

Learnings and tips from crowdsourcing project within company

Couple weeks ago I was asked to start up a crowdsourcing type of project inside the company. I was given sponsorship from our management, but I had to recruit the people myself and organize the actual work myself. I really liked the challenge and got up to speed quickly. I had only couple of weeks to finish this and I knew I had to get couple hundred people working on this on voluntary basis.

Now that the project has turned out quite well, I wanted to share my learning's from this journey to you and to analyze and learn from those myself also better. I at least analyze better written than only in my head.

Get it up and running
The first thing was to decide how much preparations I would do before really starting the project. I had basically two alternatives, (1) create proper web or tool based mechanism to divide the the work and create reporting tool or then (2) use good old excel and email based mechanism. Due to limited timeline, I selected the latter one. Within one day from announcing the project, I already had tens of people working on the actual work items.

This was a tough choice and I believe in some cases more elegant systems would help on coordinating the work and results. Of course I had a good plan also, just the tools were ancient. I juts believed that in this situation the best thing was to but the wheels running fast. That' was the way to get more time for our volunteers and fast feedback on my methods. So off I went.

Recruiting people
Like said, this was a company internal project and it has it's own challenges on recruiting people to join. I had to get sponsorship from our management (my manager did this for me, thanks for that), but I had to convince line managers around to give this a go and then convince the actual people to join to help a common goal.

Only tip I can say about this phase is, that be honest and transparent. Tell exactly how much of a time it might take from volunteer, what's the timeline and when results are expected. Also be clear what are the prerequisites people have to fulfill to join. And remember to tell the goal of the exercise clearly and understandably. Keys to get people to join are that they understand what they work for and what they exactly need to do.

Walk trough the whole chain
Before you roll up your sleeves and put the ball running; walk the whole chain trough once. What I do first, how do I instruct people, how do I get the results back, what do I do with the results, how the results are shown, who are the key people to know about the results, when this thing will finished and so forth.

Make an easy walk trough of the most important steps you expect you will face during the project. It won't take more than 15-30 minutes for this exercise, but it will definitely help you to see whole.

Test your instructions
This is the part I learned a lot. I made the perfect instructions at first. I looked the subject from many directions, answered all the questions in advance, made pictures with explanations on those and walked through the chain with my instructions. Then I asked few of my colleagues to check those. They found some minor tweaks to those, but after those, the instructions we ready to rock. Or so I thought.

I still decided to take one more cautious step before rolling the instructions to everyone. I decided to use target group for verifying those. When the first voluntaries arrived, I gave the instructions for the first tens of people and paused the roll out for a while. I asked my target group to immediately report any misunderstanding and clarification needs from my instructions. What happened was, that within the first day, I did four major updates to my thought to be perfect instructions. I congratulated myself for using this focus group approach. It turned out I really needed it.

Do not assume
In the first set of instructions I made two assumptions, that (1) some things are common knowledge and  (2) that it is obvious that I expected people to communicate back to me with certain way. Both of those turned out to be false ones.

I learned that I should never assume, that things are commonly known, because me and my close colleagues are familiar with the subject in advance. I had to make updates to my instructions about these really basic issues. I had actually explained the harder parts really well, but then the easy ones I had failed to explain well enough.

Biggest mistake of this whole project was, that I didn't specify the way people should communicate back to me. I assumed everyone would get the idea from my instructions, but it turned out to be quite the opposite. I got contacted with so many different ways and so many different formats, that I had to use lot of time on standardizing the results together. This is where the more advanced reporting tools would have become handy. I made conscious decision to stick with simple tools and get it running faster. It would have easily taken 2-3 days of my project to develop those tools, so I still think I made the right call with excel and email. Next time I just must use more time to explain how I want the results to be communicated to me.

Emphasize possibility to ask for more information
This was a learning from past similar projects. Some people need to be encouraged to ask for more info. Some people are naturally shy and they try to find information from sources they are familiar with. This easily leads to false information due to uncertainties and misunderstandings. So I encourage to put an extra emphasis on ways and possibilities to ask for clarifying questions. Tools like email and IM chat's work well for naturally shyer people.

Friends are the worst ones
I've done these type of projects couple of times in smaller scale and based of those and the project in question, I can tell that friends are the worst ones. Those people who you now quite well and who you do have a relationship with, easily don't obey the deadlines and read the instructions nearly as well enough as the ones you don't know in advance. Somehow the friendship status makes people to be more careless and they don't use the attention needed to successfully complete the tasks given. Of course there are exceptions to this and some friends can be even used for help. Overall I still think that friends are the worst ones in this type of projects. 

Have a realistic schedule
Before making a target schedule for your project, think about the scale. If there's 100 people involved, completing a project in 2 days is a big challenge. From 100 persons, there's easily ten  who come up with something more urgent to deal with and it might be hard to get replacement for so many in such short notice. I don't have any rule of thumb here, but just think of how long the task might take, double it and then give still some buffer. People always works in different ways and different situations.

When you give people a deadline, don't put it to the last day you want it to complete. Have it at least 1 or 2 days in advance and this gives you the possibility to be flexible, when people come to you to ask for more time. They are happy and you will still get your results on time.

Have a good plan of follow-up
As Stephen Covey advised in his book 7 Habits of highly effective people: "Habit 2 - Start with the end in mind", and so you should do here also. Always keep in mind what is the purpose of this all. Is the way people are working the important thing or do you care only about the results. Or are the results actually just a side note and the purpose of this all is to come up with actions from the results. Or are actually the meaning of all this to fix some underlying problem and you are just the step towards it.

So think carefully why it is done and what are the things you need to care for most. There's going to be problems in the way that something doesn't go as planned, or some results might not be as accurate as you would like those to be. Just try to reflect those to the purpose of your exercise. Like life in general, you can't win all the fights, instead try to win the war.

Written by +Henri Hämäläinen

No comments:

Post a Comment

Word is free, please leave your comment here: