How we spent 3 months and €40,000 to build our own time tracker. And why everyone should do, too
Web studios with large-scale projects will find this situation painfully familiar: Everyone’s working overtime, but efficiency is far from ideal. This is exactly where Fragment found itself in 2017. People were swamped, but billable hours, it seemed, were disappearing like water in the sand, eating away at our profits. This can happen — even to those who have been in this business for a longtime.
When we saw that there was room for improvement, we decided that for better efficiency we needed our own time tracker.
There’s already a million time trackers on the market? Why build another one?
First order of business was to put in writing all key data. How many hours were allocated, how many were used and on which tasks. We’d already been traсking our activities using Toggl (since 2013) and had all the data we needed. So we set out to putting it down in writing in the most straightforward form: in paper or in Excel.
Total budget: 100 hrs
Hours spent: 120 hrs
Balance: -20 hrs
A specialist spent 150 hours over the course of one month, including 45 hours spent on fixes or revamping studio website. Which gives us a balance of -45 hours.
It turned out that, presented this way, one a single screen and free of interpretation, the data showed just what we needed.
So, we decided to build a tracker. Who cares that there are thousands readily available! First, none of the existing solutions was ideal. Second, building one from scratch shouldn’t take an experienced team like ourselves more than a couple of days.
Or so we thought.
From makeshift utility app to finished product
Sure enough, we got the basic features, like current time budget statuses for each project and task, sorted out in just five days. There was of course the very small matter that it fell on project managers had to put in the data. As in, take the data from Toggle, then manually insert it into our table.
After two weeks, we decided we had to build an in-browser version of the tracker. Now each member of the team could fill in the hours spent on each task by pressing Start and Stop in the web app. At this stage, we had a service that met all of our studio’s needs, but we didn’t stop there, and here began the quest that ended up taking three months of our lives. But in the end we got ourselves not just a useful utility app, but a finished product, ready for marketing in the West.
Not that we had a strategy in mind, no. Just build a landing page, add registration, and upload. Why not?
It didn’t take long for us to see that at the MVP stage, we were still ages away from the finished product.
The approach we chose from the onset was minimalistic. Zero-UI design (nothing should be too visible, with the exception of the landing), no built-in interpretation — Pulse aimed at representing the data raw, thus allowing users to interpret the figures as they saw fit.
Even with such minimal specs, the design stage still turned out to be a lengthy process. Just selecting the typeface took four days. We ended up going with Typonine Sans. From there it really took off:
Interfacing — 15 days
Landing — 3 days
Front end — 12 days
Programming — 58 days
Payment processor (braintree) and invoicing — 11 days
Try application Mac, Windows — 7 days
Altogether 106 days to release, compared to the 5 days it took to build the MVP. Taking into account the fact that Scada is an experienced team, and that we had the luxury to work free of the ‘client factor’. Put in money terms, Pulse would have cost no less than EUR 40,000.
We launched in late April, 2018, when Pulse was uploaded on Producthunt, and after it was approved, Pulse was ranked #3 on Product of the Day. Other platforms followed suit, such as land-book.com, fontsinuse.com, betalist.com, and others. In terms of conversions, Producthunt remains our principal platform to this day.
From what we can see, we made the right bet with the design. During the first week, Pulse attracted 3,309 unique visitors, with a registration rate of 8.5 percent. Reviews started to come in, along with rankings of various product and landing lists. Right now, roughly 50–60 people are actively using Pulse, with a total of 1,050 registered users.
The majority of our users are designers, art directors, heads of development, and web studios. All traffic is organic, no systemic marketing effort has been undertaken, with the exception of Google retargeting ($50 per month), and several boosted posts on Facebook.
And what did that accomplish?
Pulse was developed, first and foremost, for creative teams. Since Fragment is one such team, it goes without saying that we put a lot of thought into the features. One common space (collaboration) and a Live mode (for employees) proved to be very useful for managers and directors.
The transparent structure of the reports, and the product as a whole, was intended based on the volume of data. Pulse changes drastically when it’s used by a large team sharing a single dashboard. The Project tab is especially useful as it allows to instantly see positive and negative time balance.
This mode is available to paid customers (€4 per month per user). However, our users have only just begun exploring the advantages of using Pulse for team work, so it would be early to speak of cash flow.
To us, the value of Pulse lies elsewhere.
First, building the tracker proved to be a great experiment in developing a product from the ground up. Second, we used Pulse as a chance to test-launch a product from scratch.
Second, we treat the hours that went into development and production as investment into ourselves. Yes, it would be much cheaper to just go sign up for a workshop in order to hone your skills, but this was is also infinitely more effective. Plus, the experience gained can be directly applied to work on new projects. Clients always know this and see us as a team that is capable of developing products like Pulse.