A few months ago I embarked on the journey of trying to build a new social network.
As crazy as this idea is, given the sheer complexity of the project, the fact that the competition in this space is huge, and that I have very little in the way of funding apart from my own savings, it was born of a deep frustration with the existing options.
Luckily, I have a fair amount of experience in bootstrapping new businesses, so I was willing to give it a try!
If you’ve read any modern startup books, particularly The Lean Startup by Eric Ries, you’ll probably be familiar with the concept of an MVP. Here’s a definition taken from this article here:
The minimum viable product (MVP) is the one that allows us to launch the product with the least amount of features possible so that we can learn and extract relevant information from this trial period and user interaction through a series of metrics and then act based on that data.
And here’s a very famous image to describe it:
Or here’s another useful illustration:
The idea of the lean startup cycle is BUILD-MEASURE-LEARN. And then re-iterate.
So, how to build an MVP for a social network?
Well, to begin with, I realized I needed to narrow my focus. I couldn’t just try to produce a social network that everyone in the world would use.
I needed to find a deep, narrow well!
What does this mean?
It means instead of finding a large market with a small problem, find a small market with a large (deep) problem. This way, I could focus in tightly and not spread my resources too thinly. Solve one small problem, and then broaden my focus as I grew. With limited budget and limited resources, this was essential.
Seeing as I’m a cryptocurrency and blockchain afficionado, I decided to focus my initial platform on that community. At the moment they’re spread out across a range of different social media platforms, none of which really suits their needs.
What were their needs, though? What do these people want? What problems could I solve for them that are not currently being solved by other platforms?
Without a problem to solve, I wouldn’t have a business!
Part of my initial idea for building a new social network was to build a platform that would actually show people great content about topics they were interested in (instead of bad content about things they weren’t interested in), and that might actually teach them something useful at the same time, so I decided to start off with that idea.
Plus, in the world of cryptocurrencies, there is a lot of misleading information out there, and the good information is hard to find, so I decided that a platform which solved that problem might be useful.
Enter, my first set of Twitter ads, for less than $100…
Okay, so, as you can see, not everyone thought the idea was good, but enough people seemed to be interested in the idea that I thought it was worth exploring more…
A platform like the one I was imagining was essentially a two sided marketplace. There was no point in having users who wanted to learn about something if there was nobody there to teach them. This meant getting experts on board who would be prepared to create great content.
Again, Twitter was my friend. A large number of people from the crypto community hang out on Twitter, and many of them have large followings. So, I created a survey using Surveymonkey for potential content creators who were experts in the field, and I contacted several hundred people on Twitter and asked them to fill it out.
My theory was this: if people even bothered to fill out my survey (which took about 5 minutes) then there was quite possibly an unmet need there.
Much to my excitement, quite a number of people filled out the survey, and many of them expressed a lot of interest in the project and said that they’d be happy to help create content for it and promote it. And some of these people had over 100,000 followers on Twitter!
Would anyone actually pay money for this, though? Because, without money, I wasn’t going to have a business either!
Enter, another Tweet!
With a little paid promotion, I managed to get over 400 responses:
Well, not everyone was willing to pay for it, but I thought to myself that if 30% of people on Twitter ARE willing to pay for it then they’re still pretty good numbers!
Okay then, next step. Before I could continue to explore this idea further, I needed a name. At the time I was doing these surveys, I was simply calling myself “Crypto Social Network” (it was only later that I took these screenshots — after I’d changed the name)
It was hard to find a domain name that I liked, but I came up with the following options:
Eventually I actually decided to go with www.trybe.one as my domain name, but it was useful to get some initial feedback. Don’t underestimate the importance of a name!
Okay, so with my idea partly validated, I needed to move on to an MVP…
What was the smallest possible product I could build that would help me learn about my audience and what they wanted?
A tweet, I realized, was the answer!
I’d already created a number of tweets, and from each one I was learning… So didn’t that fit the definition of an MVP?
And given that I wanted to build a social network, Twitter already had many of the functions that I would be needing, so what easier way to start building?
Plus, for any new startup, building an audience is just as important as building a product (some say far MORE important), so building my audience on Twitter definitely wasn’t going to be a waste of time.
The first thing I wanted to know about my audience was what type of content they would respond best to. I knew I wanted to teach them something, but what? Did they want to learn about startups? About ICOs? About trading cryptocurrencies? Or did they just want recent news?
Instead of creating all those tweets myself, I simply tweeted other people’s content, and then measured the response from the audience (which, by the way, was an audience of people who were interested in cryptocurrencies and blockchain technology as defined in the Twitter audience interests).
Here are the initial results:
With any test it’s important to establish which metrics are actually important to you. While the engagement rate on some tweets was high, those didn’t actually lead to people following me, which was the metric I decided was most important here.
Going by this metric, the bottom tweet was actually performing best, which is a great article on crypto trading by one of my favorite Medium authors, Daniel Jeffries. You can find it here if you’re interested.
What did this tell me? That people were most likely to really engage with high quality, long form, educational content.
Okay, so what else did my audience want?
As you can see, people wanted to be paid for their contributions, and also to be able to follow topics they were interested in (as opposed to just following individual people).
By the way, just on a side note, I found it’s a lot cheaper to advertise on Twitter if you set a “target cost”:
If you set the bid type to “automatic bid” then you’ll get a lot more clicks a lot faster, but it will cost you a lot more. Also, set your daily budget at no more than around $150. If you set it at more than that, then Twitter will make sure you fill your budget, but it won’t be so targeted in your audience — meaning you’ll end up with a lower response rate and conversion rate.
Anyway, I was now armed with some very useful information about my audience. It was time to embark on the next stage, which was to create the next version of my MVP.
In this case, a simple landing page with a signup form would serve me well.
I would describe the product in detail, and see how many people signed up to become “early adopters”. People won’t give away their email addresses that easily these days, so if people signed up to my mailing list it was my theory that they would probably sign up to my platform once it was ready.
I used WordPress, but you could use any platform for this, and came up with the following site:
To direct people to the site, I used the following tweet:
Rather than focussing on trying to get average users to sign up, I wanted to try to get content creators to sign up. I knew from my previous experiments that if I could get people to produce high quality educational content then people would engage with it. Getting people to produce this in the first place was going to be harder.
While my CTR (click through rate) was low (only 0.25%), my sign up rate was pretty good (15% of people who clicked the link and visited my site actually ended up signing up to my email list).
So maybe this was my deep well. A very small number of people, but who actually had a large problem to solve (one they felt it was worth giving away their email address for a solution to).
Okay, so this was a good start. I was building my mailing list. I was building my Twitter following.
What was next?
The next job was to continue learning about EXACTLY what my audience wanted, and trying to improve on my results. To begin with it was costing me around $50 per signup, but by the time of writing I’ve gotten that down to $25 (after just a few days), and I’m confident I can get it much lower.
My next job was to create a range of different tweets to see which one performs best.
To judge the true performance of any advertising you need to measure the CONVERSION RATE, not just the number of clicks an ad receives.
In Twitter ads, there’s a link up the top under “Tools” for “Conversion Tracking” and if you’re going to do any advertising on Twitter (or anywhere else for that matter) then I highly recommend you use it:
To set it up correctly you install the universal tracking code on all pages of your site, and then you choose one page as a “successful conversion page”. This is a page that people are directed to after they’ve done what you want them to do — in my case, sign up to my mailing list. So once people sign up, they get directed to that page automatically, and it then tracks them as a conversion.
Once I’d set up my conversion tracking, I then ran ads on a whole range of different tweets to see which one performed best.
Here are some initial results:
While the images were similar on each tweet, the tweet text was different.
I’m also going to run a range of different images and see which of those performs best:
Once I’m done, I’ll have a very good idea of which particular tweet is most effective at (a) getting people to click through to my website and (b) getting people who will actually sign up to my mailing list!
Now, the next thing I need to test is my landing page itself. It’s all well and good to work out which tweet works best, but now I need to work out what version of my landing page is going to work best in terms of getting people to sign up.
So, I created another version of this page, with the focus on a few different features. It was more or less the same as the one above, with the following section added in:
When split testing, it’s a good idea not to change too many elements at once, as you won’t know which element made the difference.
To measure the effectiveness of each version of my website I installed a great little WordPress plugin called Simple Page Tester. You just create two versions of a page and test them against each other. Again, you have to set up a tiny bit of conversion tracking code, but it’s pretty simple (just a matter of pasting some shortcode into the “success” page.
Anyway, that’s it for now! I’m going to run these tests and others for a few weeks until I work out the optimal version of both my Twitter ads and also my website landing page. By knowing what landing page people respond best to, I’ll be able to learn which features people are most interested in. And also what combinations of image/text (branding) work best.
And all this before I even start building my product!
Of course, once the platform is built, there will be an ongoing process of measuring, learning, and re-iterating, but at least to start with I’ll have a very good idea of what to build before spending too much money on it!
And that’s how you build an MVP using Twitter ads!
In the next part, I’ll be discussing how to build a more complex MVP using WordPress.
Thanks for reading.
Tom Norwood.Your Remaining Votes (within 24hrs) : 10 of 10