We created a Snapchat custom geofilter tool. Unfortunately, so did Snapchat
An insight into the development of YelloGeo — the Mint product that never was
What follows is a true story.
It’s a fast-moving tale of opportunism, entrepreneurialism, innovation and, ultimately, broken dreams set in the cut n’ thrust world of digital product development. Hold onto your Moleskines.
TL;DR: We made a Snapchat geofilter creation tool, but then Snapchat released their own just as we were about to ship. You can download the geofilter template files we were due to launch with, if you like (they’re free).
The idea
One of Mint’s esteemed front-end developers, Brad Few, is a big fan of Snapchat. Indeed his Twitter bio boldly states: “I’m better at Snapchat than you.” Brad’s obsession led him to spy an opportunity when, in February 2016, Snapchat announced that users could create and upload their own geofilters for use on the service. Having previously created a tool to allow people to embed their Snapchat name on websites, he had become very familiar with the platform and its audience.
Brad immediately clocked that most Snapchat users would have neither the artistic skill nor the patience to create and submit custom geofilters that would do special events justice. Generally speaking you don’t want a shoddy-looking geofilter for a friend’s wedding or birthday party — you want something that looks classy, that people will want to actually use.
Brad surmised that there was therefore a potential market of people who would be willing to pay for a tool that would help them create their own geofilters.
Validation
With a storied history of developing our own products, Mint has the skill and experience to pursue our own ventures — indeed we’re proud to say it’s one of the things that differentiates us. So when Brad pitched the idea to the Mint directors, they were able to see the potential and set about ‘ringfencing resource’ (i.e. they worked out who could work on the product and how much time they could spend on it).
Meanwhile, lots of people started offering what, at first glance, could have looked like competing products. Geofilters.co, geofilter.studio, eventfilters.com… virtually every conceivable possible use of the word ‘filter’ was in use as a service name and URL.
However, we felt that people using the majority of these services would be taking something of a stab in the dark, as the relationship followed a client/agency model — only, without any physical meetings to build trust. Over email you’d tell them what event you needed a geofilter for, and they would go away and design it, leaving you to trust that:
a) the geofilter would be delivered in good time
b) the quality would be worth paying for
c) they’d be willing to make tweaks if you weren’t satisfied
…. and so on. In short: they relied on customers trusting strangers about whom they knew very little.
What nobody appeared to be offering was a foolproof tool that used nicely-designed templates and an onscreen interface that allowed people to essentially build their own geofilters. So that’s what we set out to build.
Work begins
Wireframes were drawn up, illustrating that contrary to the risk of emailing a stranger, this tool — which had acquired the working name of Snapetto — would use a simple, quick and easy interface to take you from zero to geofilter hero in as little as a minute or two. We also decided that rather than building a native app, we’d save time by starting out with a web app that could be used in a browser.
Brad and designer Tom were given two days to pull together a functioning prototype that would validate the assumption that we would be able to create geofilter assets programmatically. Wireframes were created to outline how the user would move through the app. This was then applied to the functioning prototype, which used PhantomJS to create screenshots as transparent .png files. These files would be saved and used as the geofilters.
You can see the prototype here. We tested it on Snapchat on a night out for Simon’s birthday, capturing red hot ping pong action in the process.
The next day we checked Snapchat’s analytics, which were pretty promising in terms of what we could expect once the product was out there and being used.
Once we were satisfied that we were on the right track when it came to functionality, we started work on a fun and colourful design that would appeal to Snapchat’s userbase.
While we got stuck into the challenges of building out the product, we witnessed noted Snapchat fan Gary Vaynerchuck — a leading early adopter in the tech world — touting the creation of geofilters as a great business opportunity. All the signs were pointing to the fact that we had spotted a market and were moving swiftly to capitalise on it; indeed the scenario reminded us of the creation of Stickygram, when we were the first people to use Instagram’s then-new API to create a tool that enabled people to put photos on fridge magnets (Stickygram was later acquired by Photobox). However, we also remained conscious that given the pronouncements of the likes of Vaynerchuck, others would be trying to do something similar.
Also in the back of our minds was the knowledge that Snapchat themselves could choose to enter the arena at any point, and that if they did it would likely mean the end of our product. However, at this point there was no sign of them doing that, and we convinced ourselves that given Snapchat’s overall growth they had enough bigger fish to focus on.
Roadblocks
As a company that builds companies while also maintaining more traditional client/agency relationships, Mint is continuously working to balance the demands that come with each model. Snapetto’s development faced competition for resource from a number of other projects, some of which were up there with the most consuming engagements we’ve worked on in the company’s 10+ years.
The result was that the availability of the required resource that we could devote to the app’s development was rarely continuous. A couple of days’ progress would be hampered by a week during which we couldn’t get a whole lot done. It’s not a new challenge, but nevertheless it is one for which there is no solution other than unlimited resources — something very few companies enjoy.
The fact that we used this project to practice using Elixir introduced challenges as well. Back-end developers at Mint have been excited by Elixir for a while, training themselves up and attending conferences to become more familiar with it. Snapetto’s timing represented a chance to put some of what had been learned into practice, and Mint’s leadership were happy to approve the opportunity.
This decision will pay off in the long term, and indeed while working on other products recently we’ve benefited from the internal knowledge gained during this time. However, using Elixir unavoidably slowed progress somewhat, if only because of the disparity between the knowledge of back and front-end developers. If a back-ender was called away to work on something else, a front-ender wasn’t able to jump in and aid Snapetto’s progress.
A threat emerges
With Snapchat geofilters now a recognised business opportunity, it was only a matter of time before someone else attempted the same thing as us. Enter PepperFilters, which gained some good traction from the Product Hunt effect after a nifty stunt involving chasing noted entrepreneur Justin Kan through an airport.
There was no doubt PepperFilters was hitting many of the same marks as our tool. You could pick templates from a list, for a number of event types, and then edit the text that would appear in the final filter. You then paid to download your filter.
Did PepperFilters invalidate our idea? We needed to figure out whether, while it was effectively doing the same thing, our product could do it better. This resulted in lots of back-and-forth within the team. It’s not easy to be objective about something you’ve spent time on, and you can become distracted by ‘sunk costs’.
However, we felt there were a few of reasons that we should press on:
1. Speed of user experience. We felt that PepperFilters was attempting to do a bit too much. There were dozens of filter templates on offer, which on the one hand is good for the user; but on the other hand if you want to choose and create your geofilter quickly, paralysis of choice can set in and the process slows.
2. Ease of use. We were confident we could provide a nicer user interface when it came to creating a geofilter. With moveable text boxes and dozens of font boxes, again we felt Pepperfilters’ approach was a bit overcomplicated. By keeping the UI quick, simple and elegant we would provide a more pleasant experience.
3. Snapchat submission. Very few services offered to take care of submitting created geofilters to Snapchat for approval. This was understandable: it’s logistically tricky and opens up a whole lot of overhead when it comes to customer service. But we felt we could handle it, and saving customers even more time and effort would be a compelling feature of the product.
With these perceived advantages on our side, we continued development. There was certainly a recognition that we had missed ‘first mover advantage’, but with Snapchat’s huge userbase we felt there was still enough pie for multiple players.
Preparing to ship
Snapchat geofilters were now being referred to as a megatrend, which while encouraging was also concerning: the number of players was likely to continue to grow, while we were missing the boat for press coverage that comes with a new tool. Accordingly we started working on a marketing plan that centred instead around direct influencer outreach.
Meanwhile we needed to finalise the name of the product. Using Upvoter (another Mint product; give it a try if you need to make group decisions!), the wider Mint team used their votes to pick the best name. Then Brad went and chose the one he wanted anyway, thereby proving that democracy is a vicious lie. YelloGeo it was.
With the name decided it was time to refine the design, add copy to the website, take some product shots, get the T&Cs sorted, and get the thing out the door.
The nail in the coffin
And then Snapchat went and did it themselves anyway, as they had every right to do.
While it wasn’t widely publicised at the time, Brad quickly spotted Snapchat’s tool that allowed people to create their own geofilters. It did pretty much the same things YelloGeo did (it even chose the same three categories of birthdays, weddings and ‘general’ congratulations) and was pretty close to the Pepperfilters approach.
Snapchat also had the advantage of not needing to estimate an average cost for creating a geofence around locations (which is how we were approaching this feature). Whatever mysterious witchcraft they used to calculate geofence costs, obviously they had free and full access to that.
But the real kicker was that Snapchat’s tool was free to use.
At any point in YelloGeo’s life, this development would have made our product pretty much redundant — assuming you know it exists, why would you pay to use a third party tool when the native one is free?
With a heavy heart the team members agreed that there wasn’t much point shipping YelloGeo, even though we were mere days away from being able to do so.
In a way it was just as well this happened before we shipped and invested more resource in refining YelloGeo, and then a bunch more resource (and possibly money) to promote it.
It’s just a shame that the ultimate validation of our idea was the same thing that killed it off.
However, we’ll take what we learned and move onto the next venture emboldened by the fact that, but for Snapchat’s move, we would have been in a great position to capitalise on a huge market.
The End (of YelloGeo)
If you like the look of the geofilter templates we created, you can download them for free at this Dropbox link.
- Each image is available as both an .svg file and a .png file. They’re transparent, so once customised they can be submitted to Snapchat.
- To open and edit the .svg files you’ll need to use software like Adobe Illustrator or similar.
- To open and edit the .png files you can use most image editing software.
- There’s also a layered .psd file if you want to adapt the designs in Photoshop.
