[Day 10] Let’s See What Happened

[Day 10] Let’s See What Happened


A few weeks ago, I built N-ter, an online tool that lets you track job applications, in 10 days from idea to launch and documented it with a daily post. Unfortunately, I had to delay the post-launch activities and this post due to some personal circumstances, so this post comes later than promised. But: Better late than never! So here is the final day 10 post. Let’s go through what happened after the launch of N-ter.co, the things I’ve learned after launch, the actions I have taken based on early insights and what I intend to do as next steps.

 

Recap of the Launch

The last two days before the launch were mainly packed with adjustments to the user interface, usability of the tool and some bug fixing along the way. As I’ve described in an earlier post, when developing a functionality, I always focused on making it work somehow first. “Somehow” in most cases is not the way that is user-friendly or aesthetically pleasing. This is why the adjustments were necessary towards the end of the 10 days to make the tool actually usable. I was having trouble with some CSS code, which ended up being the reason why I had to delay the launch by one day and making this effectively an 11-day series.

The launch itself was quite smooth. Once the solution was in place and tested by myself, all I had to do was enable open user registration in the WordPress backend, remove the coming soon page that you would see if you visited the site before the launch and communicate the launch itself. Next to a launch post on Reddit I sent out a launch notification to the mailing list.

 

Analyzing the First Two Weeks of N-ter.co

If you’ve been following the series, you will probably remember the conversion funnel from day 3. The conversion funnel is the central tool that we are using to measure the success of our website. This is how it looked like:

  • Website visitors
  • Out of those > Signed up
  • Out of those > Logged in
  • Out of those > Created a job application
  • Out of those > Returning users
  • Out of those > Clicked an affiliate link (“Paying customer”)

 

Let’s have a look at the actual metrics from the first two weeks after launching N-ter.co:

2-Week Conversion Funnel for n-ter.co

A Closer Look at the Metrics

Before we analyse the figures, there’s one consideration which is important at this stage: Pretty much all of our visitors were coming from this post series, as I didn’t market the site on any other channels yet. This means that a large percentage of people visit the site out of curiosity how this project turned out. The figures will most likely change and should be re-evaluated once we market the tool to our actual audience.

Here’s what we can learn looking at these numbers:

  • About 1 of 10 people signed up to the site after visiting the front page.

While the figure is not too bad as a generic sign-up rate, this could be better at this stage of the project. Essentially, we were able to convince only about 10% of the people, most of which I assume are genuinely interested in the outcome of this project AND who made the effort to visit the webpage, to sign up and test the tool.

  • Almost half the people that signed up, didn’t even create a job application.

This is quite a surprising takeaway from my point of view. I would have assumed if a user is already making the effort to sign up, he would at least create one test application to see how the tool works. I’m not overly worried about this, but we should keep an eye on it further down the road.

  • Pretty much all the job applications created appear to be test applications.

Only one user created at least three job applications, which leads me to assume, that all other applications were purely for testing purposes.

  • The overall user numbers are not sufficient to ultimately validate our falsifiable hypothesis

Let’s look at the falsifiable hypothesis we created on day 3 of this series as the key hypothesis to validate with this MVP:

>> I believe that 1% of the users visiting our website will sign up, create 3 job applications and return to the website twice within the first two weeks.

The key metric in the conversion funnel we wanted to achieve the “1%” for is: “visited 3 times”. After the first two weeks, this metric was 0.63%. By just looking at the percentages, one could say that we’re not very far off the 1% mark. But the problem here is with the absolute numbers. Only a single person represents these 0.63%. My takeaway from this: The “overall” user pool of 160 users that visited our website is simply not large enough to confidently validate our assumption. In retrospect, we should have added a minimum number of visitors to achieve, before looking at the numbers.

 

What Do the Insights Mean for the Way Forward?

My initial plan for the time straight after launch was to implement job ads and affiliate links to job portals to generate revenue for the website and therewith enable the final step in the conversion funnel. In short, the focus would have been on improving the conversion in this part of the funnel:

It doesn’t make much sense to work on ads and affiliate links though, if no one is using the tool yet. Instead, I want to give the validation of the current MVP another shot by making it easier for people to find and use the tool. So, I shifted the plan to get more people to visit the website, sign up and actually give the tool a try. The goal now is to

  1. Fill the funnel (a.k.a. significantly increase the number in the “visited the website” part)
  2. Improve the conversion in this part of the funnel:

Also, I will slightly adjust our falsifiable hypothesis to include my previous thoughts:

>> I believe that of 1,000 people visiting the website, 1% of the users will sign up, create 3 job applications and return to the website twice within the first two weeks.

What I did here is to simply introduce a new threshold of 1,000 people to the equation, as I was not convinced by the validation results with a small number of users. In retrospect, I should have added this threshold straight away. We will now re-validate our MVP once we reach that number of website visitors.

To clarify, what I’m doing here is not what you would call a “Pivot” in Lean Startup terms. To do a pivot we would have to adjust a key part of our business model or MVP to test a new hypothesis. In this case, I was simply not satisfied with the metrics of the first MVP in regards to overall user numbers. This is why I’m going to try and give the validation another chance with higher user numbers to figure out whether our MVP is actually viable or not.

 

Adjusting and Marketing Our MVP for a Second Shot at Validation

So, let’s give the validation another shot. Here’s what I intended to do to reach the two goals mentioned before:

  • Market the website for more detailed validation
  • Improve the landing page
  • Communicate value of using N-ter
  • Make it easier to sign up and use the tool

What I already did in the past weeks is to give the landing page a refresh and include more information so the user gets an understanding of what value the website provides to him. Next to visual improvements, I also included a small screenshot of the new dashboard and a description text on the top of the site. To further improve the conversion rate, I also enabled social registration and login with Facebook and LinkedIn. Now there’s no need any more to enter all kinds of information, check your emails and set a password on registration. Logging in to either Facebook or LinkedIn is enough – and your profile on N-ter will be created with one click. Together these improvements should hopefully get more people to sign up once they visit our page.

You can see check out the new landing page here: https://n-ter.co

What’s left now and will be one of the next steps is to market the website appropriately to fill our conversion funnel more. This will most likely include posts on appropriate subreddits and a few other websites.

 

Recapping the 10 Day Journey

The 10 (okay, 11 😊) days I spent creating and building N-ter were some of the most stressful days I’ve ever had. It included many firsts for me, especially on a technical level, so of course there were some unforeseen roadblocks along the way. And still, I’m very happy I did this as I learned A LOT during that time. I don’t know yet whether N-ter is going to survive and if yes, where the road with it is going to lead. But I feel that no matter the outcome: The time spent building it was definitely not wasted. Here are a just a few basic things I learned during the 10 days, that I want to share with you guys:

Define exactly what you are trying to validate before building your MVP

Before sitting down and starting development on your product, clearly define what you are trying to achieve with it. Try to make your goal as “black and white” as possible. What I mean by this, at the end there should be as little room for discussion as possible. A simple “yes” or “no” answer is what we are striving for here. Because once the MVP is done, it is very easy to get emotionally attached to it and debate around the grey areas trying to make it work. You can see the best example of this right in this series, as I discovered I didn’t set any goals regarding overall user figures and am now revalidating the MVP with a new parameter.

Focus on less functionality

Ask yourself: Which 20% of the functionality is going to provide the user 80% of the value. Then only focus on making those 20% reality. It will save lots of time, but will still give you a proper MVP for your validation. When building N-ter, I struggled with this at times. Cool ideas for new functionality pop up all the time and the urge to implement them is huge. But try to always remember that you are trying to validate your product. All these cool new features will not provide any value if no one is using your product in the first place. Also, don’t be afraid to do stuff manually at first, if it saves you development time. If that functionality ends up not being required, you’ve saved valuable time. If it does, great! You can still automate down the road and now additionally have a better idea of what exactly is required, as you’ve performed the task manually a few times.

Don’t hesitate putting the product out there

For this I’m going to quote Reid Hoffmann, founder of LinkedIn:

“If you’re not embarrassed by the first version of your product, you’ve launched too late.”

As long as you’re not a big corporation with lots of reputation to lose, put your product out there as soon as possible. Discussions with early users, supporters and critics help shape your product. And even if the idea doesn’t work out in the end, at least you got some valuable feedback that you can use to make your next product better. For me personally, the online exposure of the business model and the product through this post series led to some valuable insights, that I probably would not have had if I developed everything alone, somewhere in a dark room.

 

Day 10 Summary

If you’re following along: Your MVP should now be out there and launched. Crunch the numbers and see whether your falsifiable hypothesis has been validated or not. If it was, great! Keep going! If not, think about how you can “pivot” and adjust your business model. Either adjust your existing hypothesis or come up with a new one to test. No matter whether your initial hypothesis was validated or not: Start the build, measure, learn feedback loop all over!

 

A Big Thank You, Merci, Danke, Spasibo and Gracias

Even though it’s been a few weeks since the experiment, I wanted to leave a huge thank you for the (predominantly) encouraging responses, positive feedback and constructive criticism throughout our 10-day experiment. By doing this I wanted to show you one of many ways of how you can build an idea from start to finish yourself, and of course test one of my own ideas along the way. I hope I succeeded and following this series gave you some new insights or ideas, or at least started a thought process at some stage. And maybe I was even able to motivate a few people to follow along and build their own ideas?! 😊 If you did, make sure to let me know. I would love to check it out!

Leave a Reply

Your email address will not be published. Required fields are marked *