Deciding to Fold a Startup


When I was growing up I moved a decent amount.  From the time I was born till the time I finished college my family had lived in 5 different places.  To those who grew up in the same house their whole childhood this must sound horrible, but to me it was great!  The idea of moving into a new house, getting a new room, going to a new school and making new friends was an amazing feeling.  I was never the most popular guy in school, nor the most athletic, nor the smartest, but every time I moved I got the chance to reinvent myself in a new place with new people.  It’s an oddly similar emotional comparison to leaving one startup behind and moving onto the next one.


I’ve never seemed to mind letting go of the past and keeping an eye on the future.  Maybe my brain is somehow hard wired to ignore sunk costs and evaluate a situation in a clear, rational, forward-looking manner…?  Last time I checked, I’m not the tin man, so I definitely have a heart in there somewhere.  For the sake of clarity, the concept of sunk costs, and the sunk cost fallacy, are related to people’s aversion to wasting resources.  These resources are most commonly time and money.  An example of the irrational behavior of the sunk cost fallacy is going to see a movie that after 30 minutes you realize is horrible, but you stay and watch the movie, to “get your money’s worth” since you’ve already paid.


This ‘sunk cost fallacy’ is extremely relevant for entrepreneurs.  We invest so much time, money and emotion into our businesses, that when things don’t go well we are inclined to continue ‘throwing good money after bad’ because we’ve already invested so much.  This is a dangerous trap to fall in to because it blinds you from the reality of your situation.


Obviously not everyone is comfortable with change, or at least as comfortable as I appear to be, but it sure comes in handy as an entrepreneur!  After 16 months of blood (from the occasional paper cut), sweat (I like working out to clear my head), and tears (only when we made our first sale), oh, and not to mention late nights, long days, endless phone calls, and occasional nausea from staring at the screen of my Macbook, I made the oddly easy decision to fold my most recent startup, ‘’.


First, let me briefly explain what Wedkey was.  Wedkey was a website that helped brides save time and money discovering and booking wedding vendors. The short pitch to investors was ‘Our products cater directly to a bride’s wedding planning process and allow them to systematically save money on key wedding purchases. Wedkey streamlines the vendor search process through vetting and curation, adds transparency to a complex and unclear market, and is making wedding planning more social.’  The reality was Wedkey was a deal site for people planning their wedding that wanted to be much more.


The Wedkey business concept started with an idea and a dream.  Capture the niche wedding ‘deal’ market that Groupon and LivingSocial were missing, create an awesome product that would be the future of ‘deals’ because the way they were being structured by these two deal giants was generally a bad value proposition for the vendor.  Early in our process we did some consumer research, but didn’t quite know what we were doing, and listened only for the answers we wanted to hear. Sadly, this was the first and only time we did research until right before we shut down.


So, in October 2012, after 18 months of blood, etc., etc., there I was, oddly comfortable and confident in pulling the plug on my baby.  Wedkey was my first ‘real’ startup; my partners and I brought it from an idea to a full product, even generating some early sales. We were a top-5 finalist company at NYU’s prestigious Entrepreneur’s Challenge, and we were invited to pitch at some local startup competitions like Distilled Intelligence. I had just spent the last 6 months working on this company full-time while my partners both still had other jobs, and it definitely wasn’t an easy decision, but it was a confident one.


What ultimately led to the decision to fold was data, some of it objective, and some subjective, but all equally valuable.  It was data I wish I had had, or had paid attention to months before.  I got so caught up trying to do what I was ‘supposed to’ for sales, marketing, product development, and operations, I never left myself the time to do what I needed to, which was analyze how my business was doing.  I ended up with a mediocre product, a great looking website, some users, and almost no customers, and until October 2012, I had no idea why.  So here’s what happened:


At the beginning we were excited just to have an email list.  We were new to the wedding industry but had a decent network of friends and family who were either engaged, or were friends with people who were engaged in the NY and Chicago areas (our two pilot cities).  We growth hacked our way to a 3,000 person email list mostly by collecting email addresses from engaged friends and friends of friends through Facebook.  We came up with our design concept, built a splash website to collect Sign Up info, launched the damn thing and were really excited when people actually signed up for ‘early access’, which really meant nothing because we wanted people using the site as quickly as possible.  We got some initial feedback from beta sign ups which forced us to create a new design concept, tested it, built out the full site and launched it with ~10 deals in each city from a variety of vendors.  Then the excitement ended.


Mistake #1: we designed our site in a vacuum without doing any consumer testing or getting feedback from potential users.

I would have done differently: I was not my company’s target consumer, so coming up with the UI and UX for our website without user feedback was a bad idea.  I became so concerned with doing things fast for Wedkey, that I missed a bunch of steps.  It was like I tried to put together a big entertainment center from Ikea without looking at the directions, and I ended up putting part Z in backwards and totally forgot to screw in part A.  We should have developed wireframes of the full UI using something like Balsamiq, then we could have gotten feedback from our target audience by either emailing people on our wait list with questions, or recruiting them for some usability tests on a platform like or We could then have then iterated on the feedback, developed more detailed page designs and gone through the process again.  While this sounds like a lot of work, it would have been hours spent at the beginning of our project that paid dividends later on.


In the 10 months after we launched the full site, we increased the number of providers offering deals from 5 to over 50.  In those same 10 months, we generated a total of $300 in net revenue on 4 sales.  No, I didn’t forget a ‘k’, that was $300, not $300k.  We were targeting a very passionate and engaged consumer (no pun intended): 20-35 year old females in NYC and Chicago, but, our member sign up volume was abysmal, averaging about 10 new members per week.  Our conversion from members to customers was obviously much worse than that.


Mistake #2: other than the number of new sign ups and dollar sales, we weren’t tracking any other site or lead generation activity.

I would have done differently: This is honestly a bit embarrassing because these basic analytics are startup 101 stuff.  Even more embarrassing, I had an analytical background (investment banking and then corporate strategy) but was new to web startups.  I had no clue how to use Google Analytics (GA), and didn’t take the time to learn it… that’s so sad.  GA has a huge resource center of both documentation and videos that literally walk you through how to get started using the platform so you can be smart about your business (here is a good video to get started). At a minimum, I should have spent a few days watching these videos, educating myself on the platform and setting up some goals, events, and accompanying funnels.  Goals in GA enable you to measure how often a specific action is taken on your website via URL tracking, for example, how many users land on the final page of your ‘Sign Up’ process.  Events are similar to Goals, except Events track the click of a specific asset on your website, i.e. a button or a hyperlink.  Similar to the Goal example, an event could track how many users click the ‘Sign Up’ button on your website’s homepage.  Side note: Events are much more flexible within the GA framework for tracking user engagement and analysis.  Some examples of Goals and Events I would have set up are:

  • Goal – Reaching the Sign Up thank you page
  • Goal – Reaching a vendor’s page
  • Goal – Reaching deal checkout page
  • Event – Sending a ‘refer a friend’ invite
  • Event – Clicking Facebook like button
  • Event – Saving checklist of wedding items purchased to date
  • Event – Clicking each of our category search buttons (cake, photographer, etc.)
  • Event – Watching the “how it works” video on our homepage


Funnels and Goal Flow reports enable you to track the paths users take to complete an Event or a Goal.  For example, I could have set up a Goal Flow for reaching a vendor’s page and GA would show me the different ways users reached that page, and where they went after they viewed it.  Similarly, I could have set up a funnel for my deal checkout process.  This funnel would have shown me how many people enter the funnel by clicking a ‘Purchase’ button on a vendor’s page, then how many people drop out of the funnel at different steps of the checkout process before completing it.


As a lead generation site, our biggest fear was a consumer discovering a provider on our site, calling the provider and negotiating a deal on their own therefore bypassing Wedkey.  We didn’t give ourselves any way of measuring how many phone calls we were generating to any given vendor, but there are some amazing tools like Twilio that would have been perfect.  Twilio enables websites to generate unique trackable phone numbers via their API for very reasonable prices (I think it’s around $2/month/number).  We should have assigned every vendor on the site a unique phone number and tracked our lead gen volume.


Two things that I have come to love more recently are KISSmetrics and Google Ventures’ ‘Startup Lab’ youtube channel.  I’ve done a full implementation of KISSmetrics here at ServiceAlley and I refer to the Startup Lab videos on a regular basis to freshen up my skills on more complex web analytics.  For reference, KISSmetrics is a ‘person based analytics platform’.  In normal person language, that means it allows me to track the site activity of every individual person that visits, and when they log in to the site, their email address becomes associated with their unique tracking ID on KISSmetrics.  This allows me to do really cool stuff like track cohorts of people across time, or identify really engaged users who I can then reach out to individually via email.


To drive brand awareness and increase sign ups, we did a bunch of Facebook advertising for Wedkey.  We thought we were being sophisticated and tested upwards of 40 variations of ad text and images to ‘optimize our click through rates’.  We had some ads direct to our Facebook page to generate Likes, and others to our Homepage to generate sign ups.  Unfortunately, we had no analytics, we had no funnels, all we knew was that we were spending money on ads, and we weren’t generating any revenue! Anecdotally we decided that Facebook was a good way to drive some brand awareness, but not necessarily instigate purchase.


Mistake #3: we weren’t paying any attention to the conversion rate of our ads.

I would have done differently: We expected that just because we were running ads, people would start to love our business.  Oh how wrong we were.  The CTRs on our Facebook ads were amazingly low, even for Facebook standards.  We had a Facebook ad specialist help us with our ads to try and improve the engagement but that didn’t help either.  It is probably obvious at this point, but we should have started with a smaller number of ads, tested some very specific copy and images, and concentrated more on what people were doing after they clicked on the ad than the actual ad click itself.  Setting up advanced segments in GA is super easy and we could have tracked just our Facebook referral traffic and how those users engaged with the site.  We also should have tested Google AdWords.  The main difference between Facebook ads and Google ads is the intent to purchase that’s inherent when someone does a Google search.  When people are looking around on Facebook, chances are they are doing it casually, not looking to purchase an item.  Hence we were trying to instigate purchase activity out of a passive audience.  Facebook is great for brand building and awareness, but is a much tougher nut to crack for commerce, and that’s not something anyone should overlook.


The straw that broke the camel’s back (my back), were the results of the consumer research interviews we conducted in September and October of 2012.  I sent out a simple invite to people who had signed up for our email list to have a 15-minute conversation with me in exchange for a $10 Amazon gift card.  I got an overwhelming interest from people and quickly learned why.  The pattern became brutally clear after just the first 5 calls… Our users loved the concept of a deal site specifically for weddings, they loved that the deals were up on our site for 30 days at a time, they loved the design, but there weren’t enough vendors for them to choose from, and most importantly the deals weren’t good enough.  We had made a key assumption early on that by vetting vendors for quality and reputation, when we put them on our site it would reduce the need for brides to shop around.  We therefore wouldn’t need a huge stock of vendors in each category (cake, photographer, etc.); instead we’d just need a few really good ones.  WRONG.  Another one of our core assumptions that was annihilated with info from these interviews was the value proposition we created for our deals.  By that I mean, wedding vendors are not restaurants; there is a very low likelihood of repeat business so every sale from our site needed to be a marginally profitable one.  In our effort to develop the ‘deal model of the future’, we structured our deals so that the vendor had a profitable sale, the bride saved ~15% and we made a little along the way.  What brides told us was that since we were a ‘deal site’, they expected discounts of 40% or more.  That just wasn’t going to fly with our wedding vendors, or with any wedding vendor… So there we were, two of our core assumptions shot to hell, an explanation of why we were generating sign ups and no sales, and all it took was $200 and a few 30-minute telephone conversations.


Mistake #4: we didn’t sufficiently vet our core customer assumptions from the start.

I would have done differently:  Started with better consumer research with a goal of not just uncovering a need, but understanding how to create a solution.  Then continuing the research constantly to ensure we were fulfilling that need.


I love and hate the phrase ‘hindsight is always 20/20’.  I hate it because everyone says it when they mess something up.  I love it because it’s fortunately so very true, and it’s an extremely valuable concept if you make use of it.  To me the phrase means taking the time to look back, reflect on what went right, what went wrong, and learning from it all to make myself a better entrepreneur, a better partner, a better leader, and a better person in the future.  In an industry I was not familiar with, building a product that wasn’t for me, it should have all been about research and understanding.  It sure would have saved a lot of time, headache, and money.


In some of my next few posts I’ll go a bit deeper into consumer research and analytics. Let me know if there is something in particular you’d like me to discuss.

Did you fold your startup?  What led you to your decision to fold?  How did you deal with it emotionally?



One Response to Deciding to Fold a Startup

  1. Pingback: Guide to Consumer Research Interviews - 6 Essential Steps | Beckwords

Leave a Reply

Please use your real name instead of you company name or keyword spam.