After leaving audible on March 2017, I took a break for three months and then joined Fiverr on July 2017. I've been a buyer at Fiverr since 2012 when I started working on a side project called upendo and have also used it on other services. So this was an opportunity for me to work on a product that I love and was very excited to join the team.
Fiverr is the biggest freelancers market and it has 10 different main categories and 130 sub categories to offer. The main office is in Tel Aviv with branches in San Francisco, Miami, and New York. The New York office handles category experience and I led the product design of it with the help of another two product designers who were reporting directly to me.
I led the Fiverr NYC design team that oversees category experience for both buyers and sellers. I worked with 3 PMs, 1 product lead and 2 designers who were my direct reports. I reported directly to CTO who is overseeing both tech and product. I did both hands-on design work and also managed my direct report’s work.
Since Fiverr is the biggest online freelancers market, there are thousands of news sellers joining the site everyday and they are the lifeblood of what we called the gig economy. All these new sellers are opening up thousands of gigs on the site and it's become harder and harder for buyers to differentiate between gigs. This started to result in a huge drop off rate for buyers. Narrowing this funnel became an important work for our team.
There are main categories like graphic and design, programming and tech, etc., that are very different at the top level, but in the sub category level, web design sub category sounds like it can be inside both categories. People who work in the industry might be able to differentiate well but a cupcake shop owner or a chef or a car workshop owner might not. So fixing the navigation to help these buyers find what they need is a good problem to solve.
Entrepreneurs with startups have to work double or triple the amount of work most people in 9-5 day jobs have, since they are in an urgent state to kickstart the work, and to prove to investors that their idea is valid. Even if they see something they like, the sellers need to be available to reply back to the buyers and start working immediately. Creating a feature and a flow to match buyers with urgent needs and sellers who are available at that moment would benefit everyone.
My PM and I went through different flows of buyers and tried to identify all the pain points. We noticed that not being able to get to the right gigs, not being able to differentiate between gigs, and urgent or on demand service to get the work done are three major problems in their flows. We decided to fix in phases and prioritized them by impact to our buyers, our workload, and future thinking. The first thing we could fix easily is by pointing our buyers to the right sellers or gigs.
A few data I picked up while doing a research is that filters can really help narrow down the results and get to a really good match, but the usage is only around 5%. I thought about creating a flow where we ask questions to buyers and answering them will turn on filters to give more relevant results. It's almost like a step by step wizard. My pm and I worked on those wizard flows to finalize the copies and what type of filters to turn on. I picked those up to create a design and in some cases trying to divide up the questions to create much more fluid experience.
I made the wizard to be big and disruptive to the screen but at the same time, I made the visual design to feel like it belongs on the page. I utilized the wizard's extra space to explain the options and their benefits inside the wizard on hover. I put ways for buyers to move back and forth between steps inside wizard to get better control. At the end, I started putting I don't know option inside the wizard flow to understand where we might need to improve.
I put in processing animation to give buyers the feeling of progress and add text to explain these are personalized.
I replaced all the logo samples to be Fiverr so that it would be easier for buyers to understand and compare how their brands could look like.
Wizard was launched incrementally to see the results. First time buyer went up by 5%, Average sale price went up by 10 percent, cancellation rate went down by 6 percent and overall usage of wizard is 40% comparing to 3% on filters.
The next thing I fixed in the flow was decision indicators on the gig cards. Ambiguity inside categories like programming and tech is just so high, even using wizard and filters still wouldn't explain the difference between the sellers.
So I started thinking about what type of decision indicators I would need to differentiate sellers and there are things like sellers expertise, message reply timeframe, sellers location and the most successful price range. But these can't be fit into a card so I started exploring list view. But creating a view is a big effort and I needed to make sure it's beneficial to other categories.
This is when my pm and I started looking into all the top sub categories and tried to decide which ones would be beneficial with current grid cards and which ones would be with the new list view. We called these experiencable categories and abstract categories. We also started to noticed that inside abstract categories, people clicked on seller profile name and Avatar to go directly to seller profile to learn more from there and communicate to the seller directly. We moved the collecting cta heart to a more prominent place since we know that people who collect tend to buy more.
This gave us enough confidence to push forward with the list view. We launched the list view on Jan 2018 and the results were amazing. In organic growth, cancellation rate went down by 10 percent, revenue up by 10 percent, average sales price went up by 10 percent.
Since we moved the favorite button out to a more prominent position on both desktop and on mobile, collection went up 15% and due to that with newly sign up first time buyers went up by 5%.
Below are mockups of the list view on desktop and mobile.
As I have said previously, most of our customers are entrepreneurs and most of the time, they work alone. So, pushing out ideas into realities is harder when you are alone. Fiverr is a place where freelancers can help them with what they need, fast. But we want to push this to a new level. We want to create a product where sellers can opt in to say they are ready and a badge will be on their gigs to let buyers know these sellers can immediately reply messages to them and immediately start working with them.
For the design process, I divided into buyer experience and seller experience. For buyers, they will need a clear indicators when they are searching or browsing, and a way to filter results only to sellers who are available now. The badges need to be on gig cards and gig detail pages.
For sellers, it's a bit more complicated since it's an opt in / out experience. I created a place on the top header so that it's easier for them to opt in and out. Once clicked a popover modal will appear with information about available now, the criterias to meet the requirements, and toggle to opt in an out.
One of the big challenges we knew we would face was to make sellers reply back to buyers when they receive messages. So we made sure sellers agree to reply back messages in less than 5 min and if they don't, we would automatically opt them out. If that happens for three times, we would disable the feature from the seller for a month.
Initial results from the release of this test were not up to our expectations. Average sales price went down, and the quality of the deliveries were low. It was basically due to flood of poor quality sellers with low cost gigs. This made us rethink how we select sellers for this test. And the fact that the toggle was put next to Fiverr pro toggle made the team uncomfortable since pro is supposed to be our best quality workforce. So we started tweaking the algorithm to get better quality sellers, and at the same time removed the toggle and created a highlighted row just for available now sellers.
With these changes, we made a huge positive impact to average sales price, cancellation rate, and best of all, quality. The highlighted row design made the buyers focused more on those gigs and buy a third more than any other gigs on higher and better positions.
Fiverr is revolutionizing the way people work all over the world. No one needs to quit their day job or need to have a day job. Somewhere around the world, there’s someone who’s in need of your help and Fiverr is making the gap smaller and smaller each day and I am proud to be a part of the team.
For confidentiality reasons, I have taken out the actual numbers from all the companies I have worked for and have put percentage. These are all just my views on the project I worked on and does not represent the views on the organizations.
© www.yairpike.com 2011-2021