After writing my forum post a week ago, and reading a lot of threads here on IH, I found that more insight into what other people are doing is not only fun, it's also useful and can be really motivational. So in true 'sharing is caring' fashion, I thought I'd share the experience of my current project. I'll be detailed about my reasoning in terms of target audience, marketplace, budget, current status and more.
It always starts with that 'aha' moment
About a year ago I had an idea for a product for internal communications. I had recently listened to an episode of the podcast 99% invisible about "H day", the day that Sweden went from left-hand traffic to right-hand traffic and the successful information campaign that led up to it. I had also attended a meeting with the CEO of a game development studio in my city with almost 500 employees. They had problems with employees not reading the internal news and information despite using all kinds of distribution channels - even using digital displays with information in the office elevator and staircases.
Being generally interested in communication and having worked with or close to marketing led me to think about this. At the time I was CEO of a small retail and e-commerce company with around 20 employees. Previously I had co-founded a web agency that had peaked at around 25 employees - and I knew that it could be difficult to get information across to everyone. Especially if people worked part time, didn't have a company e-mail adress or didn't work in front of computers. We used an internal newsletter via e-mail, but I still heard a lot of "I didn't know that" or "no one told me about this". So I figured there had to be a better way to do this. A way that's set-it-and-forget-it for managers.
Thus the idea for Trench was born.
The first few steps
I had initially thought of creating a tool that used all possible channels at once: Publish to intranets, Slack channels, digital signage etc. I even wanted it to be able to automatically create a printed news digest that got sent to employees homes as well. But after a lot of brainstorming and discussions and having to actually become more specific in terms of features, I settled on a slimmer version.
At this time I also plotted down my target audience, companies with the following traits:
- At least 25 employees
- Staff that work uneven hours, part time, geographically spread or similar. I.e. not in the same place, meaning that physical staff meetings are difficult to manage.
The reasoning behind this is that there's a limit at ~20 people where it gets more difficult to reach everyone with information. And if people work remote or part time in a retail store chain etc. it gets even more difficult.
Below are two matrixes I did to figure out for myself where Trench made sense compared to other ways of communication as well as other tools on the market.
Validating the problem
I knew that it's very easy to get ahead of one self when it comes to product ideas in general. You get swept away by dreamy and unrealistic visions of the perfect future of the idea. I wasn't going to do that. I wasn't going to fall in love with my idea, alteast not yet. I needed to validate that the problem existed first.
Ask and you shall receive
I had recently read a really good article on the topic of validation and was eager to get started with that method, i.e. calling companies and asking them about their problems and their priorities. I was travelling around in India at the time so phone calls were out of the question, so I opted to make a survey instead.
Since I've done business for a while I had a network of contacts to use, so I composed a survey using Google Forms and sent it out to a selection of contacts. The questions I asked were quite broad, and were about spreading news or information within organizations. Most people I got to answer were either managers, HR people or CEOs. People that worked with spreading news and information internally.
I got around 30 replies, and the data was rather interesting. Most companies rated the effectiveness of their own internal news flow as a 3 out of 5. Not low enough to validate the idea perhaps, but a lot of them mentioned that they lacked good tools, and that news easily "slipped away" and got buried in peoples e-mail inboxes, intranet notifications or Slack channels. Something that did surprise me a little was that a lot fewer companies than I thought had intranets, and that almost all of them used e-mail newsletters or plain "one off" e-mails for internal news.
All in all, the data I had gathered wasn't enought for me to feel that the idea was either valid or invalid. I needed more.
Fake it 'til you make it
I knew that a lot of people make "fake" landing pages in order to see if there's any interest for a product. I knew that the method was very useful and had tried it twice before with other ideas - and had managed to invalidate them (which can be a really good thing). I don't think its usable for all business ideas, but for an SaaS I figured it was worth a try.
I quickly wrote some copy, thought out a design concept I wanted to use and started building a simple landing page:
Then I set up an Adwords campaign targeting keywords such as "internal newsletter", "intranet staff news", "employee newsletter" etc. I had one ad group aimed at people looking for intranet solutions, another aimed at newsletters for staff and a third one for more general search terms related to the product.
The first version of the landing page was not very effective. At all. I used Hotjar to get a better idea on how visitors scanned the site. I was most likely too vague in my message and the copy was way too long. This wasn't really news to me since I've worked with web design for about 10 years, but sometimes you forget to use your own knowledge properly. I started A/B testing the site with a new design, that's still there now:
The new design with smaller features and a better description of the product worked, perhaps not surprisingly, way better. I got some pretty cheap feedback using Usersthink which, albeit not crucial, helped me improve the design a little as well. I also had some friends (that I knew would be critical) pitch in.
I also had to tweak the Adwords campaign quite a lot during the 2-3 weeks I ran my test. I got a lot of junk visitors looking for logins for their own company intranet. Apparently there's a market for setting up fake "American Airlines Intranet" login pages and running AdSense on them. After a couple alterations I got it running smoothly though, and with what looked to me like real traffic.
I spent about $500 in total and had a 5% conversion rate for my "early access" signup. The first $200 went straight down the AdSense-fake-intranet-login rabbit hole though. I also added a pricing table to the site about halfway through the test which didn't have any negative impact on the signup rate.
All in all I had a rough estimate of $20 dollars per signup that I could use as a base for my CAC and CLV calculations. Even though a signup is not the same as a paying customer, having some real data is way better than no data and all when doing a budget. Speaking of budgets..
Why budgeting matters
Around 2011 I spotted what I figured was a solid market opportunity. There was only a single software company in the realty market in Sweden. They owned the market and bought any company that contested them. The software itself was technically simple: Realtors could upload photos of a house/apartment and post the object to an online third party marketplace. But it was expensive and a first generation software of this kind, i.e. a Windows application instead of being web-based. On the surface it looked simple: Build a competing software with a lower price point and sweep the market. But, when I did the numbers, I realized that there were actually way less registered realtors in the country that I had thought - meaning that the numbers simply didn't add up - and I threw the idea in the trash.
To date, atleast three other companies have attempted to go after that same market, and none of them have succeeded.
Sure - there's Facebook and what not that didn't spend time trying to make money until they had a big market share. But there's a reason they're called unicorns, and I wouldn't bet my own money on an early stage company like that.
Doing financial calculations might not be everyones' favorite task, but in my opinion it's crucial to any business idea. Otherwise you might spend months or even years building something that isn't financially viable that you could've easily simulated if you would've just spent a few hours on it.
Our minds aren't naturally good at complex calculations - so taking your time doing this is always worthwhile. It doesn't have to be a sophisticated budget but it's important to at least try to gather some data and do some calculations. It makes things much more tangible. Is $X per month really a viable price point for your service? How many customers would you need, and at which growth and churn rate? Is there a big enough market? What are you basing your data on? Remember that there's often a lot of government or industry statistics available that could be of use.
I had run a rather successful landing page trial along with a survey - both indicating that I might have had a solid idea for a product. I also had some initial calculations indicating that the business could be financially viable. But I wanted more. I knew that I had cheated a little by being on the lazy side in regards to reaching out to potential customers, and that I was probably applying some confirmation bias to my survey results. So - I decided to call up some of the survey participants that I knew in one way or another and ask them a little more in-depth about their problems.
The result I got was in no way definitive (I don't think it ever is either), but it did seem to me that all of the managers I talked to genuinely expressed that being able to keep their staff up to date and informed was a top 3 priority. They also lacked good tools for the job.
Now - I didn't talk to them about my solution at this stage because it wasn't set in stone. I had some different ideas on how the product would work and I didn't want to decide on a feature set too early. I wanted to be able to keep an open mind and get a better understanding of their problems. I also in no way wanted their opinion on what the product should be able to do. The reason for this is that a lot of people have ideas and want to contribute, but starting that discussion will quickly turn it into a rather useless "ooh, what if it could do this and that" situation.
At this point I had to make a decision. Do I:
- Scrap the idea and move on with my life
- Continue trying to validate the idea
- Move on to the next phase and build an MVP
I knew that it's almost impossible to get a solid validation without having an actual product, no matter how much research you do. I also felt that I had too much validation to not do something with it. So, I decided on the latter, building an MVP.
Minimum viable product
I have a lot of experience from a lot of different fields within web design. I made my first website when I was eleven or twelve (Netscape Composer FTW), and kept it as a hobby during my teens (Remember Kirupa and "photoshop tuts"-sites anyone?). When I was 24 I founded a web agency with a friend that we grew steadily into a 20-something employee operation. During those years I did a little of everything there is to do in an agency: from marketing to fullstack development and project management. I've also done a fair share of design, both print and web. This means that I'm capable to do most of the things needed for an MVP myself, but doing everything from design and development to marketing and planning in a single project is not easy, especially not when you're doing work as a consultant in the meantime, pouring quality time and effort into client projects.
So I decided to get external help.
The first thing I did was write a short functional specification for the project. This is not necessarily needed when you're alone in a project because it's easy to keep everything in your own head. But I know how important it is to convey your idea properly to other developers, and since I had decided to hire an offshore developer I had to do it. It's a good exercise either way, because having to be more specific is actually very useful as a way to get out of the comfort zone and be truly specific about features that might have been pretty cozy, vague and defendable in your own mind.
With the specification for the MVP in hand, I posted the job on Upwork and had a few short interviews with a little less than 10 developers. I knew what to look for since I've vetted developers before when hiring for the agency. I found a talented freelance developer in Russia that I got along with really well and decided to hire him. The plan was to get the MVP built within 3 months with me working about 25% of a work week on the project and him doing 75% work weeks. Work started in mid july and has been ongoing since - but working on the MVP led me to my next obstacle.
Getting pilot customers
I knew early on that I had to get at least two or three customers from my target audience in at an early stage. There were two local companies that I had relations to that I knew from the start that I could get on board for the MVP. However, none of them were in my target group, and weren't in any real need of the product, so that wouldn't have covered it. I also wanted testers that I didn't know. Companies that truly needed the software and whose motivation for testing the software was for it to solve their problems with internal communication.
I also had a meeting with a Swedish government organisation that supports startups. They normally only do lending which wasn't relevant for me, but they had a smaller program which featured an initial grant of $5.000 and potentially up to an additional $10.000 for subsequent stages. My meeting with them had gone very well, so well that they wanted me on the program. They only had one request: Get a couple of LOIs (Letters of intent) signed, showing that you have customers willing to pay for your software. The money was very relevant to me, as it could cover almost half of my total expenses for the MVP, with the rest coming out of my own pocket.
At this point, as mentioned in my forum post I tried reaching out to a couple of companies that I had interviewed in the initial phase. It didn't go well. I discovered the hard way that people that initially are glad to help with answering questions are not necessarily as happy to sign up for a beta of a software they are not sure they need.
This shouldn't have come as a surprise to me, but in hindsight I should have been more aggressive during the interview and research stage and maybe pushed more buttons.
At this point I had an MVP that was about 50% done. Completing it wouldn't be a problem, but it would cost me additional money to do, and with no one in my target audience to test it yet, I had to find other ways.
The word cold is not very attractive. We associated it with things that are negative. So it's no surprise that cold calls or cold e-mails isn't something that most people enjoy doing. I'm no different, especially not as the introvert that I am. But - I knew I had to do it, and so I chose the path with the least resistance for me: cold emails.
I purchased a list of e-mail addresses and company details belonging to approx. 150 HR managers in my city (Malmö, Sweden). I wanted them to be local because my product wasn't ready, and I didn't have a link to send them that would get them hooked. An actual physical meeting means so much more so that's what my aim was with the e-mail. The first e-mail I sent out looked something like this (albeit in Swedish):
I sent these out in four different batches over 5 days, tweaking the content a little for each one. A couple of days after having sent out the first wave of e-mails the response looked something like this:
- Response rate: ~10%
- Out of office: ~4%
The response rate didn't seem bad, but the responses were all negative. This wasn't unexpected, but it didn't help me with my motivation.
From white walkers to thaw
I wasn't about to crawl into fetal position just yet. So I sent out a second e-mail to everyone that hadn't responded yet:
This time I got a bigger response. I got an immediate sense that everyone that had read my first e-mail but postponed a response were much quicker this time around. After a few days had passed the stats looked like this:
- Response rate: ~25%
- Out of office: ~6%
I did learn a little from the negative responses, because I replied back to them, and managed to get some information from them in regards to the tools they use to day and why they weren't interested. These were the answers I got:
- Interested but not right now, get back to me in X weeks/months
- Happy with current solutions (being mainly physichal meetings or e-mail newsletter)
- Small company, didn't relate to having these problems
But - most important of all: I got two meetings. Funny enough, both of them used my link to book a meeting on another slot than the one I had suggested. I wonder If they would've responded at all if I hadn't made things dead simple for them.
And this is where I am right now. I have those two presentation meetings booked in the upcoming week, along with a meeting with an HR manager from a gaming company in my city who agreed to let me pick her brain. I have also paused all development on the MVP because I don't want to spend more time or money on it if I can't get any users at this stage.
I will most likely decide within two weeks whether to kill my darling or continue the struggle. To be continued!
Some numbers for those of you who'd like to know..
I think it's important to keep track of all expenses when doing something like this, both your own time and actual money spent. We humans are so bad at estimating stuff like this, so if you don't do it from the start you'll have problems making proper estimates when looking back at something you've done months ago.
I've tracked every minute I've spent on this project, because I know that it's difficult to know what the true cost is otherwise. This is my current status:
- Survey 3 hours
- Landing page 30 hours
- Adwords 4 hours
- Validation Interviews 3 hours
- Recruiting offshore developer 8 hours
- MVP spec and development 40 hours
- Getting pilot customers 10 hours
- Budget, matrix, research 8 hours
- Various meetings 4 hours
All in all about 110 hours.
- Offshore developer ~$3000
- Adwords ~$500
- Various tools ~$200
- Purchased e-mail list ~$200
All in all, something along the lines of $4000
Now - someone might argue: "Why spend money on another developer when you can do it yourself?". First of all, I'm more of a generalist than a specialist and not an awesome backend developer. Second of all, if the project takes off, I don't want to be the only one with a sense of ownership over the codebase. And third: I make a lot more money per hour with my consulting work than it costs me to hire a developer, so it's actually both better and cheaper for me to do it this way.
Relevant services or tools that I've used so far
- Calculations and documents: Google Docs/Sheets
- Calendar appointments: Calendly
- Hiring freelancer: Upwork
- Time tracking: Toggl
- Advertisement for the landing page: Google Adwords
- Recording user behavior on the landing site: Hotjar
- E-mail campaign: Polymail Pro
- Landing page review: UsersThink