Wysker Treasure Hunt

6 min readFeb 15, 2021

Back in the Summer of 2018, I became involved in the wysker treasure hunt. As wysker put it:

To promote the wys token, we didn’t just want to create simple banner ads that spread out all the trending keywords: crypto, coins, Lambos, moon, hodl, whale, fud, pump and dump. So we thought, how to do something different? How to create something fun and challenging? That’s why we created the wysker treasure hunt.

The idea was, by solving clues, to obtain the private key for a digital wallet containing 100,000 wys tokens. The wys token was a utility token, it could initially only be used within the application.

In a nutshell, the wysker business case was to create a shopping app with which users could earn tokens to get discounts on products. Users would sell their shopping data to advertisers to earn tokens. In turn, the advertisers could better target their products to customers. The token thus provided a direct incentive to spend time in the app.

What I’d like to discuss is how to measure the success of a treasure hunt as an alternative marketing strategy. I will describe the setup and solution of the wysker treasure hunt, after that I will analysis the treasure hunt in form of a classic marketing campaign.

Treasure Hunt Setup

The treasure hunt centred around the launch video. The video was also provided in original MP4 format, in addition to 977 individual images of the entire video. A second advertising video would also prove to be part of the treasure hunt.

Communication tools were a Telegram group where seekers could exchange hints, tips and tricks. A Reddit community where seekers could interact directly with the wysker team and a Medium post to kick the whole thing off.

Hints where distributed on Telegram and Reddit on a semi-regular basis to keep seekers engaged. The hints were essential to solving the puzzle, so they could be used to guide the overall time span of the treasure hunt.

Solving the Treasure Hunt

There is a more detailed writeup available, I’ll just quickly go over the solutions (not in order):

Tiny letters in the corners of individual frames of the launch video provided the first part of the solution.
QR-Code hidden in a frame that could only be found by modifying the contrast and brightness of the frame. The URL encoded in the QR-Code lead to a piece of the key.
Audio from the video contained samples that when slowed down and played backwards, lead to an email address. Sending an email to that address, revealed another piece of the key.
A further piece of the puzzle was found in the SVG logo of wysker. As a slightly off-colour layer in the SVG.
Based on a quote and a specific frame, needed to send an artists name to an email, to obtain yet another piece of the key.
A final piece was encoded as morse code in the time square video.

Interestingly the original MP4 video contained metadata which allowed seekers to short-circuit the treasure hunt. Whether this was intentional is left to speculation.

So quite a bit of effort went into this treasure hunt. The use of different technologies i.e. audio, image, video, QR-Code, etc, and hints that referenced TV series or Carl Sagan quotes made it engaging and — literally — fun.

Strangely the installation of the application was not necessary to obtain any puzzle pieces. The only person who would have needed to install the app would have been the winner, to spend the tokens.

It took approximately three-and-a-half months for the hunt to be solved. Whether this was expected or not is something that only wysker knows.

In terms of a marketing campaign, what were the relevant numbers for the treasure hunt? How could success and progress of the treasure be measured?

Analytics and Marketing Funnel

How do you measure success for this type of marketing campaign? Tracking pixel and cookies are of limited use for this treasure hunt. Existing tools are of only limited assistance for analytics. What could be measured instead?

Medium claps (~ 160)
Size/Activity Telegram group (no. of members > 1,300)
Reddit activity (no. of members 119)
YouTube views (> 120,000)
Website views
App Installs

None of these provided a complete picture of the user journey through the funnel — the funnel being the discovery of puzzle pieces. There seemed to be no clear approach to computing the funnel conversion rate.

However three pieces of the puzzle could be tracked directly. One was a link and two other pieces were obtained by sending emails. So for those puzzle pieces, a mini-funnel could be established.

A fourth piece was in the SVG logo but that’s part of the website, so it would be difficult to differentiate traffic and to know which seeker had discovered that piece.

The other two pieces where directly embedded in the videos, so there was no direct indications on how many seekers had found those pieces. In direct, monitoring the discussion in the Telegram group could have provided clues.

Funnel conversion rate was then reduced to only three puzzle pieces. Hence it wasn’t possible to get a complete segmentation of how many seekers had how many pieces of the puzzle.

Seeker Segmentation and Hint Timing

For the most part, releasing hints was the only instrument available to control how fast seekers went through the puzzle, and thus the time frame of the entire treasure hunt.

And without having a complete segmentation of seekers into groups of who found what, it’s difficult to decide on the timing of new hints. Releasing hints too soon might discourage those seekers who weren’t that far, releasing hints too late might not reach those seekers that had lost interest because they found all pieces up to the hint.

Target Audience

Each marketing campaign has a target audience. Ideally target audience should be end users of the product who would end up monetising. In this case, the treasure hunt was aimed at promoting the wysker token. But to whom?

Was it to promote the token to Ethereum investors or end users of the application?

I would argue that the technical savvy needed to solve the treasure hunt made the seekers not the same as the end-users of the application. In addition, since the installation of the app wasn’t necessary to solve the treasure hunt, a seeker wasn’t even exposed to the application.

This probably caused a disjunct in seekers that were engaged in the treasure hunt and those users that were target audience for the application. Could this have been corrected during the campaign?

Campaign Controls

It’s a very static form of marketing campaign since all the puzzle pieces were there from the start. This meant refocusing the campaign to a different target audience during the campaign would have been difficult.

In this case, the only adjustment possible was in the release of the hints. All the piece of the puzzles where already out there. However, there was nothing stopping wysker from duplicating puzzle pieces and changing future hints to point to the duplicated pieces. For example, so as to hide the last piece of the puzzle in the application, thus making the installation of the application necessary.

A early stop to a treasure hunt campaign can only be done by announcing a winner. So a treasure hunt must be done until someone solves it, there is no real possibility of cancelling it earlier without a winner.

Conclusion

I would argue that the wysker treasure hunt, as a marketing campaign, missed its goals. The target audience seemed to be misplaced and since the analytics and campaign controls weren’t on point, it was hard to refocus the campaign to another target audience.

As a marketing campaign, this treasure hunt was definitely something new and that made it fun. Of course, there is direct correlation between difficulty and fun, and difficulty has a direct correlation with effort. That is to say, setting up a treasure hunt is quite a bit more effort than a classic banner campaign.

Making puzzle pieces only available via trackable sources, e.g., links or emails, makes user segmentation possible and the analytics are well defined. Of course, these numbers will always be a base minimum value since seekers will share/swap puzzle pieces, skewing the analytics.

An alternative approach would have been to release puzzle pieces with the hints. This would allow a realignment of the target audience but would probably have increased the effort required.

In the end, all respect to the wysker team, it was great experience going through the treasure hunt!

Thanks for reading!