I recently posted an ad on Kijiji to sell light-medium data projects, like scraping public datasets from websites, R development, or helping with Excel & PowerPivot. These were the data science and data management skills I thought I could use to help others. My mentors advised local communities were great starting points to offer these services so I decided to try it myself.
I thought it was reasonable to anticipate that a few independent retail shops, restaurants, or home businesses were curious about unrealized insights within sales data accumulated through the years. It was also plausible they’ve been wondering if they can learn from FANG stock companies‘ data strategies to better their own business.
After several weeks, I sold my first data project. The client was a financial services professional with a tailored list of stock symbols that needed several data points to be appended to it.
The following are my key takeaways from this experience:
This wasn’t my client’s first data rodeo
While I thought my target market was centered around business owners looking to adopt more data and technology into their growth strategy, my first client was actually looking for another means to collect the same data. They already had a solution in place that collected stock data.
Transparency and flexibility were some issues I heard as reasons why my client started window shopping for other solutions. This really echoed the importance of being transparent in today’s information age, where consumers can easily access information about many products and services. Designing a reasonably flexible product, while maintaining a level of simplicity so clients understand it, can also be a challenging balancing act that could make or break a deal.
The delivery was as much work as the development
A lot more time was put into testing and exceptions handling than when I was coding data collection scripts for my own projects.
When I run scripts to collect and organize my own data it wasn’t detrimental to have to rectify and work around fatal errors as I go. However a product sold had to be debugged and running error-free to ensure it was a useful piece of software for my client. It was designed to be reused as often as desired, as oppose to a snapshot of the domain (a single data set).
In addition to ensuring the program didn’t crash, I had to be cognoscente the client might not have Python installed on the machine doing the web scraping. In this case, I made a trip to their office to install Python and all necessary third party modules, plus a quick tutorial to get them up and running.
I needed to lead with the value my project will bring
Years ago in my twenties, I was commuting across the city about 60-min one-way through gridlock traffic. From this, I had an idea put companies’ ads on my car for a fee. I then proceeded to register a business name, a tax registration number, and got to work on market price research and posting on as many online classified sites as I could. It took at least 12-18 months to finally close a deal. It took that long because I didn’t lead with value.
There were many inquiries in those 12-18 months, which all broke down at the pricing negotiations. I declined several requests for free-trial periods when in hindsight, a free-trial period could have delivered some value first. If the ad campaign was productive for the client, it’s hard to imagine they would have stopped that value stream.
This time around, I focused most, if not all, my attention on delivering what the client wanted from the get go – a solution to missing data points for stock symbols. Fees were not discussed until I felt the client was starting to conceptualize some value out of their work with me. Referring prospects to my previous data project articles certainly helped sell myself too.
I was focusing on delivering solutions to problems first, putting customers’ and prospects’ needs in the forefront.
This was a humble win to start. There’s more markets to explore as my first customer already came from a different motivation than I thought I was targeting. To anticipate unexpected requests, familiarize with more tools to keep step with the general demand in the market. Needless to say, if a request is recurring, go learn, practice and offer it.
For myself, it would be great added value to deliver a Python or R script that can run without having Python or R installed on the machine. How to deliver a solution via live code publishing web apps such as Zeppelin Notebook would be great to learn.
I’m looking forward to what’s next in another stint for me with an entrepreneurial venture and will share more success stories your way.
Thank you for reading!