Our previous blog post about GoDutch.cash talked about our first unsuccessful attempt at launching a product, and how we learned that it’s important to talk to users first, and not over-build before releasing. So what did we do next? We tried to make another product and made a lot of the same mistakes again!
To be fair, the timeline explains that a little bit. We started building GoDutch more than a year ago for our own purposes, and started work on HospoSafe about a month ago. In the meantime, through discussions with other aspiring businesses, reading a lot online, and also speaking with readers of this blog, we’ve gained a much clearer understanding of how we should be approaching these projects.
Nevertheless, we now know we’ll need to be a lot more strategic in the future.
What is HospoSafe?
HospoSafe is an app designed to help cafés and restaurants to comply with COVID-19 contact tracing regulations in Victoria and New South Wales, Australia. Hospitality businesses need to record a contact name and phone number for all visitors, store this data securely for 28 days, then destroy it. A lot of businesses are doing this using pens and paper.
HospoSafe gives businesses a unique check-in URL and generates PDF posters to print out with a QR code. Customers scan the code at the door, fill out a simple web form, and show a green “successfully checked in” screen to staff members. The data is stored securely, and purged after 28 days, but can be downloaded in the interim if government tracing teams need the info.
About a month ago, things were looking up here in Melbourne.1 COVID-19 restrictions were easing and restaurants, cafés and bars were opening up again. We had heard about the new contact tracing requirements for hospitality businesses and thought that a self check-in app could be a great idea.
We searched the web and found a handful of competitors. Most of them were free, but they looked pretty janky2 and weren’t clear about security or how they were handling private data. There was one very slick competitor, but they were very expensive, and it seemed that you needed to request a call from their sales team before you could even sign up. So we thought there was a market, and we could differentiate ourselves as a fellow Melbourne small business, offering a privacy and security-focussed, lower-priced option that people could start using straight away.
We set ourselves the deadline to launch an MVP by 1st July, giving us about a week to build it. And we did! 🎉
But, we then had to grapple again with the question of how to get eyes on the product. It seems like a silly oversight now, but we still hadn’t really thought that part through. We don’t have any direct connections with restaurant owners. The thought of cold-calling businesses made us cringe, but we were OK with the idea of sending emails or Facebook messages.
We were also starting to read more about marketing and landing page optimisation, and we realised our landing page could use a lot of work. We decided to start with Google Ads again while we were improving the landing page and planning how we would reach out to businesses directly (we might have been procrastinating a bit here!).
The reality check
As all this was happening, community infection rates of COVID-19 in Melbourne were increasing. Lockdowns were re-imposed on parts of Melbourne on 2nd of July, and the whole city a week later. Restaurants and cafés were back to serving take-away food only.
Our original idea was to market ourselves in Melbourne as a fellow local small business. Restaurants in regional Victoria and other states were still open, so we weren’t completely deterred by the lockdown.
But doing more customer research finally gave us a reality check. We found a Facebook group offering support and advice for local hospitality business owners during the pandemic. As we scrolled back, we found other competitors that we had never seen before, including a slick-looking one built by university students who were offering it for free.3 We also found posts from a month ago where restaurant owners were complaining about being inundated with sales calls about contact tracing apps.
Infuriatingly, a lot of the offerings were explicitly suggesting restauranteurs could use the contact details collected for exposure tracing, as a marketing tool. Apart from the underhanded ethics, this is disallowed by the guidelines. But what can you do – the temptation to use customer contact details for marketing is probably too great to resist.4
What we tried
We spent $68.11 on Google ads, which got us 34 clicks and one person signing up for a free trial.
Was it a success?
No. We had one sign-up to HospoSafe from a pub owner in regional Victoria via Google Ads, but it looks like they decided not to use it, and they didn’t respond to our email checking in on how they were going.
We’ve thought about trying to promote the app to restaurants outside of Melbourne, but at this stage we think we were probably too late to the party with HospoSafe.
We still weren’t thorough enough about research on the competition or validating demand before we started building. We should have made more effort to try to find the places online where our users go, before building anything. If we had found that Facebook group back in June, we might have had a chance to make connections and find potential users before starting to code, or we might have still concluded that it was too late, and not wasted time building an app.
We spent too much time messing with CSS. We used the Bulma CSS framework, which is a helpful tool, but we ended up arranging columns and buttons by hand, and the app still looked more janky than we would have liked. Using a UI kit with pre-made UI components would have saved us valuable time and given a more polished result. If anyone has recommendations for this we’d love to hear them!
Next time, we should first find a community of potential users, such as a Facebook group or a Slack channel, where we can gather input before building a product.
As usual, we’re battling the feeling that we wasted time and effort, but we must remind ourselves that we would approach it quite differently if we were to start now. That means we’ve learned something, which is worthwhile in and of itself.
One of them even required the restaurant patron to search for the establishment by name, and the others looked like 2002-era websites. ↩
Unfortunately, while it looked slick, their app was actually fairly broken. The download-a-QR-code feature just didn’t work. Of course, we have no idea how many users they actually had. ↩
While the government instructions don’t literally forbid you to use the contact details for marketing (“Visitor and patron contact details should be shared only with representatives of the Department of Health and Human Services for the purpose of contact tracing. Additional information should not be collected for privacy reasons.”), this clearly isn’t what they had in mind. ↩