3 Part Series On Startup Failure: Part 3, Validation


Customer discovery and validation

Cultural Differences Spawn Opportunity

In the summer of 2007, I had the opportunity to move to Berlin with a large company. Relocation was part of the package, which meant helping me with visa and work permit issues, finding housing, and even getting started with the language. Moving was a very pleasant experience, and my new home felt as comfortable as the one I had left in California.

Fast forward a couple of years and I could fill out a book with the cultural differences that make Germany a very different place than the US. I previously wrote about 3 Ways Americans Can Learn From Germans, but in that article I didn’t mention that I actually launched a business around a major cultural difference between Germans and Americans: if you can pay someone to make your life easier in the US, there’s a demand for that service; in Germany, it almost felt like the cultural norm was to do things the hard way, just because there was an outdated rule that said that’s how it should be done and any deviation from that outdated rule would be treason. That’s probably not a totally fair assessment, but it’s how I, and many other foreigners, felt with certain aspects of German life. If interested, you can read a bit more on how to be German in this hilarious blog post by a Briton in Germany.

Customer Service For All

With this background, in the winter of 2012, I decided I needed a change in my professional life and wanted to draw upon my years of international management, MBA, and cultural experience, to launch a bilingual personal assistant service to make life easier in Germany. I imagined a comfortable blanket-service that would serve as a buffer between foreigners and the often confusing circumstances of every day life in Germany. The first version of Ask Geoffrey was born, with the original robot:


The business concept was simple: customers could email in their tasks, questions, or problems, and a local bilingual assistant would help them resolve it. Each request was limited to 15min of assistant help; bigger requests would need to be broken up. The service would be provided on a subscription basis (X amount per month for Y number of requests).

Collecting Interest

I tried to be smart about the business, following the validation approach of gauging demand before fully launching the service. So I setup a simple landing page through Launchrock, and started telling friends about the service, tweeting links to signup, and posting on Facebook. The list of submitted emails started to grow. So I decided to open up shop for the first small batch of users and start processing requests myself. It worked, and people liked it! They referred more people, and the email list grew to a few hundred.

Running Lean

At this point, I decided that this system would only grow if I figured out an efficient way to contract local assistants, train them, and get them running in the system. (The “system” was a cleverly hacked-together system of emails and off-the-shelf customer service software). So I did exactly that: prepared job postings, interview procedures, & training material. Before long, I had 2 freelancers in the system answering ad-hoc requests and getting paid per request.

Turnaround time on requests was really important, and the freelancers were not always immediately available. So I thought it would be wise to increase the number of freelancers in the system to increase the chance that one would be quickly available to answer new requests. I assumed the incentive of getting paid per request would encourage them to be first to answer.

Users – Yes; Customers – No

Over the course of about 5 months, users (free users, careful distinction from paying customers) steadily grew to over 200, with over 100 of them active on a monthly basis. The system was working ok for the most part, with occasional screw ups on requests, or extended delays in response times. In July 2013, I introduced the pricing model:

  • €25 per month for 5 requests
  • €45 per month for 10 requests
  • €85 per month for 20 requests

Of the 200+ registered users, about 3-5 signed up over the first 2 weeks. BURN. I decided it would be a good idea to offer discounts, so I dropped the rates by about 20%. I got a few more signups, but nothing near 200.

At the same time, I was in discussions with some newfound friends in Tel Aviv who loved the idea and wanted to try it out there. We spent a good amount of time mapping out the strategy and translating the website. In the end, it never happened. Sidetrack and distraction.

Meanwhile, back in Berlin, I thought the lack of interest was due to a lack of publicity. So I went to work trying to spread the word – through personal and extended networks. I even managed to get a few articles (in German and English) written about the service in blogs or local tech news. It barely moved the needle.

At this stage I had poured a few thousand Euros into the website (nice design matters for B2C products, screenshot of landing page below), payout to agents, and admin fees for setting up a legal entity. But the actual (revenue) numbers were nowhere near the estimates in the financial plan I had so carefully crafted.

Pivoting from B2C to B2B

Around Q4 2013, I shifted gears to reposition the service as a perk for employees that businesses could offer, because I finally learned that B2C is a very challenging model and B2B was a potentially far better model for this service, in-line with my original subscription model. Furthermore, we had a couple of power users for business purposes and paid for by their companies, and these were the best customers in terms of accountable revenue as well as low maintenance.

By the end of 2013, it was clear that the service would need more investment and time to reposition as a B2B service and it came down to personal belief in the service. And while I loved building the business, I had to admit that I was not personally fulfilled with what I was building. It went from trying to fill the void in customer service culture in Germany, to taking care of largely trivial tasks for the international elite. Perhaps the future may have been bright, but in order to pull myself through the rough patches, I would have needed the burning belief that this service was truly solving a far-reaching problem. And I had to be honest with myself that I didn’t believe it anymore. I shut down in January 2014, transitioned customers to an alternative service, closed the Berlin chapter of my life and moved back to California in February 2014.

Five Lessons Learned

  1. Lack of validation: I thought I was validating the business, but I wasn’t really because validation includes the service/product at a given price; free users are obviously different than paying customers, but I was too caught up in the idea to realize this.
  2. Didn’t address or frame a problem: we offered the service as a convenience to individuals who didn’t necessarily need it, or didn’t have the perception that they needed it. This links back to validation, where we should have done a far better job at identifying the real burning pain that we could have solved and focused on that key topic.
  3. Wrong market: my network was in the startup scene in Berlin, which is vibrant, but not necessarily well-financed. And having positioned the service as a convenience, most were quick to categorize it as a nice-to-have instead of a must-have problem solver. Again, we should have found the most urgent pain we could solve and focused on that market.
  4. Lack of focus: we tried to solve too many problems for too many people. While that was our eventual goal, our service offering was too broad and general so people didn’t make an immediate connection between pain and solution. We should have focused on singled out pains, established a loyal base, gained credibility, then expanded to other pain points we could address well.
  5. Complexity: at one point, we had 10 simultaneously contracted agents. This happened with the customer in mind and making certain responses were provided within 1-2 hours max. Instead, it had the opposite effect where agents were demotivated as it meant a lower chance of a significant earning for the month. For example, instead of paying €20 to 10 agents, we should have concentrated €200 on 1 agent.

So What?

Writing this 3-part series was a great exercise in reflection. We tend to get caught up in the fast-moving environment of our work and life, and it really helps to pause and take a step back once in a while. As much as I hope I don’t get caught in the same mistakes and failures that I’ve written about, I’m sure I’ll relapse to old behaviors once in a while, but I also hope I’ll catch myself in time to correct course. The one mistake I’m clear about now is to pursue an idea I care about. As many have said before, startups and the entrepreneurial lifestyle are an emotional roller coaster; there’s no challenge in staying the course when everything (or at least most things) is going well, but the real test of your perseverance is when everything seems to be going wrong – what will keep you motivated then? Most likely a strong support system of friends and family, and your own conviction that what you’re doing is worth the (temporary) pain.

Thanks for reading. You can also read the first 2 parts on Execution and Focus. Since moving back to California, I’ve started working on the first marketplace for medical writing. Check back again next month for a post I’m working on about the US healthcare system, a very broken system that is in desperate need of reform.

Screenshot of the Geoffrey landing page (click and zoom in to enlarge):



3 Part Series On Startup Failure: Part 2, Execution



2008: The Markets Collapse, An Opportunity Emerges

In the Fall of 2008, I was sitting in Berlin, glued to the BBC TV station, watching hourly reports of the NYSE dropping at record levels. This was the unraveling of the US financial system that eventually spread like a malignant cancer throughout the global financial community. It was surreal.

During this time, my soon-to-be-cofounder of our second business together was working at a commercial real estate consulting firm in San Francisco. Their firm was booming by taking advantage of California Proposition 13, which enabled owners to calculate property tax based on current market value. This was especially useful during the post-2008 financial meltdown, as many people had purchased homes at the peak of the real estate bubble pre-2008, and were paying property tax on a higher value than they should have been.

For example, if I had bought a house in early 2008 for $700k, I was probably paying about $7k per year (1% of value) in property tax. But in 2009, when real estate went south, my property dropped in value to $500k, meaning I was paying an excess of $2k for property tax. Prop 13 gave me the right to file for a property value re-assessment to current market value, and correspondingly adjust my annual property tax to $5k.

tax reduction

Automating A Manual Process

Re-assessing property values was usually done by consultants who would charge large fees (upwards of thousands of dollars). We figured out a way to automate the process using Zillow API’s and publicly available county forms. An online turnkey solution would pull comparable values from Zillow, walk you through the form, and provide you all the documents needed for the appeal process in PDF format to send into your local county office. The entire package would cost about $50 and save you many multiples more. It was a home run idea in a high-demand market, all it needed was a robust website.

Too Little Planning, Too Much Outsourcing

We basically knew what we needed on the website, and had to develop a first prototype to get in front of customers and test it out. Neither of us was able to setup a dynamic website pulling data from Zillow, so we had to hire a third party. Our budget was limited and we found an outsourced solution in India to complete the project. They spec’d out the project, provided exactly what we asked for on paper, and we agreed on milestone payments in 4 parts.

At the first checkpoint, we had basic functionality. At the second checkpoint, we made further progress in terms of function and some aesthetics. At checkpoints 3 & 4 we started to recognize our cultural differences and extreme level of micromanagement needed to create a logical UX and decent UI. We lost our patience and instead of staying focused, working together with the dev team to complete our vision, we grew frustrated and lashed out at the project manager for being incompetent. We grew increasingly frustrated with the entire situation and felt that we had come too far to turn back. We also realized our contract was setup in a way that we had to complete all 4 payments in order to take ownership of the code. We thought we could take what they had done and hire a freelancer to polish it off. Unfortunately, once we did this and showed the code base to other developers, they explained that the entire project would be better off scrapped and started from scratch as the code base was a total mismatch with our goals.

Three months and thousands of dollars later, we were out of budget and time. Competitors started to pop up and we lost our window. Our golden opportunity had passed us by. It’s still frustrating to think about it today; we definitely learned our lesson the hard way.

Four Lessons Learned

  1. Whether you’re hiring a marketeer or a programmer, make sure his/ her skills match the job requirements, which also means you need to know (or have access to someone who knows) about that subject.
  2. Map out the user experience as much as possible before submitting anything to code. Use power point or any other means to walk through the entire process, show it to friends and strangers, and make sure it makes sense. It’s your job to understand your customer’s needs.
  3. Stay focused on your end goal, especially at the hardest of times. It’s easy to get things done when they’re going well, but it’s critical to stay focused when everything seems to be going wrong.
  4. Allocate a realistic budget, and then set aside another 50% as a rainy day fund. You don’t want to get halfway through you development, only to realize you under-budgeted and won’t be able to deliver the best possible product.

Thanks for reading. In the next and final installment in this 3-part series, I’ll review a business that had a proven business model but failed at the most crucial step, validation. You can also access the first part on Focus here.

3 Part Series on Startup Failure: Part 1, Focus



I graduated, now what?

In 2001, with a fresh degree in Molecular Cell Biology,  I was lucky to find a relevant job at a pharmaceutical company where I used at least 50% of my college education. About 1 year into it, I grew weary of the slow pace of development and change in my work environment. Even though I was doing exactly what I had studied and loved, there was something missing. I wasn’t challenged enough. I wanted to do more. I wanted to move faster. But that’s not how a large corporation is setup, at least not in R&D. 

Bored at work? Take the first (safe) step

So I got together with a group of friends from various industries ranging from real estate, to construction, to finance (who coincidentally felt the same as I did), and we started a small round table discussion once a week on potential business ideas we could start on the side. Our discussions varied all over the place as we each evaluated our strengths and access to markets/ products. One idea was a commercial water filtration business, another was a “we-do-anything” marketplace/ hotline (kind of like TaskRabbit), and the final one we landed on was a tutoring company.

Focus on your strengths, advantages, and expertise

Why tutoring? We all attended good high schools in an area where families place high value on education, and we all attended good universities. We had access to both: top-notch university students (supply), and high performing high schools with well-off families invested in their children’s future (demand). We knew we had stark competition (Kaplan, Princeton, other local companies), but we also knew that if we could provide a unique selling point and excellent quality, we could make a difference.

Making mistakes is part of the process

This was our first business launch, and as such, we hit the usual snags of team-member fallouts, administration before validation, and generally slow progress. We did get very lucky in that we invested in administrative issues (business cards, flyers, registration) far before landing our first client. While we didn’t test anything systematically, we did gather lots of feedback from friends and family. We also paid a lot of attention to design, an early version of our logo is below (we liked the frog as it is an adaptive species to many environments, hence our company name as well).

Screen Shot 2014-01-04 at 11.01.20 AM

Disqualify suppliers and customers, focus your market

Next, we decided to put the supply side together, the actual tutors. We had access to UC Berkeley and St. Mary’s college as our main local sources, so we hit the ground and posted to job boards anywhere and everywhere we were allowed to do so. We wanted to recruit top quality, so we offered good pay for the time, $18 – $20 per hour. Pretty soon, we started getting applications and we screened and “disqualified” applicants with a simple word document application that also required letters of recommendation and a professor vouching for the topic they wanted to tutor. This required some effort on the applicant side, but then provided quality leads. Check out the original application here.

With tutors on deck, we developed our first flyer with lots of feedback gathered from friends and family. We went in person to the local high schools we wanted to target and spoke with administration and career counselors so they would get to know us personally and hopefully recommend us to their parents and students. We placed our flyers in their offices, right next to Kaplan’s and Princeton’s. As you can see, we didn’t provide pricing on our flyers, instead we offered a free consultation. This enabled us to capture warm leads and sell manually until we built up our reputation.

Flier copy

I don’t quite recall every detail of our first sale, but I will never forget how awesome it felt. High fives were definitely exchanged afterwards. From there, word of mouth fueled our growth and we steadily grew our client base and started earning some revenue.

Trying to do too much, too fast

It was about this time that the unraveling started. We were delivering great service (in-home tutoring) at a decent price (just below the market leader’s rates) and building lasting relationships with the students and parents. Instead of focusing on building a product out of our strength, which was basic academic tutoring and relationship management, we wanted to do more, such as SAT prep and college counseling, because those margins were fatter. We clearly did not have the expertise in those areas, and thus began our partnership development struggles with successful individuals who were not necessarily a good product fit. Our product was a 2-way partnership between the student and the academic tutor, resulting in an environment of academic accountability, while SAT and college prep services provided subject expertise and a curriculum to follow. As we spent more time negotiating partnership agreements, we lost focus on our core service and growth stagnated.

A divided team with unbalanced work

Meanwhile at real work, my company offered me a short time assignment overseas in Germany. I saw it as a great opportunity to do something different and I accepted the offer, assuming I will be able to continue my tutor and client management role from overseas via emails and Skype calls. Once I moved, everything unraveled as our processes were highly customized so it wasn’t easy to simply shift responsibilities, and it was even harder to maintain those responsibilities with a 9 hour time difference. So basically I spent several weeks up late at night into the morning on Skype calls to manage tutor schedules, last-minute changes, and client relations. My business partner ended up overloaded with work and we were unable to continue our operations. That Fall, about 2 years after launching, we shut the service down and passed on our clients and tutors to a friend who was running a similar business in a neighboring region.

It was an awesome experience that taught me what it meant to be an entrepreneur, and I loved it.

6 lessons I learned

  1. Focus on 1 thing at a time.
  2. Validate demand (multiple successful competitors is a good indicator).
  3. Talk with people to understand their specific needs, and tailor a product to meet those needs better than available solutions.
  4. Establish a USP that significantly differentiates you from competitors (guaranteed quality and results is usually a good place to start, albeit very difficult to deliver).
  5. Manually setting up and running a business teaches you all the ins and outs, but remember to think about growth and processes/ automations to facilitate that growth.
  6. Work on something you love and believe in, education checked those boxes for me.

In the next installment, I’ll review a business that had tremendous potential, but failed on execution.

Please provide learnings from your failures in the comments section below, or get in touch with me directly.

Thanks for reading, and thanks to my co-founder David for reviewing this post.