How we used Engineering As Marketing To Get More Sign-Ups

I first read about Engineering As Marketing in one of the posts by Gabriel Weingberg, Founder & CEO of DuckDuckGo. It sounded like an interesting concept but I didn’t think about it much. We were too busy doing pre-launch marketing the conventional way: creating content and promoting it.

Our arsenal of marketing tactics also included answering quora questions (answers got 12k views in total), cold emailing (how we got those emails deserves a separate post of its own) and optimizing site layout to get higher conversion from whatever traffic we were getting.

Initially we were a mobile first, email based CRM for Photographers. Writing a pitch for an explainer video forced us to self-reflect and we realized that our primary goal all along was to create an A.I. app rather than a conventional CURD application.

So we did a mini pivot and ditched Photographers for Self-Employed Techies as our beachhead market. The hypothesis is that it will be easier to lure techies into becoming early adopters by promoting the Machine Learning aspect of our product. After the pivot, we scraped our existing marketing strategy and decided to try something new.

Zuckerberg’s Hacker’s way became a source of inspiration. We decided to apply the philosophy of “move fast and break things” to marketing. The plan was to tryout a new marketing “hack” every week and see what sticks and what doesn’t.

So here is a list of everything that we have tried after our last pivot. (click on the underlined headings to view experiments)

Experiment # 1: What the F is ML?

Inspiration: DuckDuckGo’s donttrack.us was successful in generating traffic for their main site.

Hypothesis: If we create a memefied guide to Machine Learning as a stand alone website we will be able to redirect some of the traffic to caspy.com and get more sign ups.

What Happened: It took us about a week of content creation and tinkering to dumb down Machine Learning and present it in a memefied way. As soon as we launched on Product Hunt we got featured. As expected, we saw a spike in traffic to our main site and got a bunch of sign ups.

Conclusion: Micro-sites are an excellent way of increasing brand awareness and doing marketing in a subtle way. We will do more of it in future.

Experiment # 2: Content with a Video TLDR

Inspiration: NowThis videos are trending on Facebook. They compress big, complex stories into few animated images and text with a slick soundtrack.

Hypothesis: If we create something similar to NowThis video animations and use it as a video TLDR to our blog post, we will go viral and get more sign ups.

What Happened: The blog post and the accompanying video took a week to create and the cost also included an Adobe CC subscription. We used Photoshop to edit images and After Effects to animate them and we got a free soundtrack from Youtube.

After launch we did notice more reads compared to previous posts but didn’t get crazy traffic. We were successful at starting a conversation on reddit.com/r/art but that could be attributed more to the title and content of the blog than the video.

Conclusions: We did see a 23% increase in first day traffic compared to previous posts but the concept still needs more experimenting.

Experiment # 3: Animated Storytelling

Inspiration: I came across Hardbound on Product Hunt and got hooked on their storytelling method.

Hypothesis: Static content is being phased out by the recent surge in gifs and animated videos. Harbound’s way of storytelling coupled with our content will helps us use the elements of the Hooked cycle to increase user engagement and get more signups.

What Happened: The animation part took longer than planned and after launch we weren’t able to get the buzz we expected. We feel that the content could have been improved.

Conclusion: There should be no trade-off when it comes to the quality of your content. The experiment wasn’t a complete failure though. We still managed to get 12% more first day reads than average which is enough to justify giving this medium another shot.

Experiment # 4: The Donald Test

Inspiration: Sortable, our neighbors, created a successful PR campaign around the 2016 US Elections. There’s nothing else in the press than the 2016 US Election!

Hypothesis: If we created an app that uses our Machine Learning Engine to tell if a person sounds more like Hillary or Trump on Twitter, we will go viral and get more signups.

What Happened: We got featured again on Product Hunt. At one point, we were getting 300 hits per second which caused the app to break down. There was a lot of firefighting but we still managed to remain on top in Product Hunt’s Artificial Intelligence, Politics and Funny category.

Conclusion: A lot more traffic and a lot more signups. We also got featured on Business Insider in addition to few other blogs. It showed us that we can piggyback on the US Election once again to generate more traffic.