As part of our ongoing marketing efforts at oDesk, we have a pretty substantial Google Adwords campaign set up. I was reviewing some of the data today and we noticed that for all campaigns, the conversion ratios (% of visitors that sign up for an account) were highest on Saturdays. A little unexpected, but an interesting finding. Google allows you to tailor your campaigns by automatically adjusting your bid amounts by day of week. Seems like it would be appropriate to increase our bid amounts on Saturdays so we capitalize on this increased conversion ratio. Or maybe not?
I was wondering why conversion ratio was higher on the weekend so I decided to take a look at oDesk data for job postings over the last few months. I took a look at over 10,000 jobs posted and analyzed results based on a few simple criteria.
First, I wanted to understand the distribution of types of jobs posted by day. I put together all the results in Excel and did a pivot table on the day of week. Note that for all of the following charts, 1 = Monday, 7 = Sunday. This first chart shows the total ratio of fixed price jobs posted on the oDesk network by the day of week. Sundays apparently have the highest percentage of fixed price jobs.
Ok, so clearly there is a small bias toward fixed price jobs on the weekends. What about job size? The following chart shows the average estimated hours for hourly jobs (estimated hours = estimated workload in hours per week * estimated duration in weeks). Note that durations over 100 weeks are rounded down to 100 weeks. Clearly, hourly jobs tend to be smallest on Sundays.
Ok, we’re starting to see a trend here. Let’s take a look at Fixed Price jobs. I looked at a chart of Fixed Price job size by day of week posted. You can see in this next chart that the budgets tend to be smaller than the average (the average being 100%) on Friday thru Sunday.
So, jobs that are posted on the weekend tend to be smaller and more urgent than the rest of the jobs posted. Now, remember that the conversion ratio was highest on Saturdays. The way I read this data is that if you’re looking for urgent work, you’re more willing to try anything to get the job done so you’ll convert to a user at a higher rate than normal. Now the last step is to look at the relative value of a customer that signs up for our network by day of week. If the “weekend warrior” types do indeed have small, urgent jobs then they should have a lower lifetime value than the average.
While there is not a huge difference, I think the findings are pretty interesting. Note for providers -> if you’re looking for the best jobs available – make sure you’re looking Mon – Fri.