Scaled and Failed #30 – A Recap 🎉

Looking Back to Learn

Welcome to Scaled and Failed! My name’s Amil Naik and I’m an aspiring VC and founder at The University of Texas at Austin currently working at Decibel. I write about startups that scaled and startups that failed to draw insights about the patterns of startup failure and how to avoid them. Everything is clearer in hindsight, so it’s worth looking back.

If you haven’t subscribed already, do it here:

Looking Back on 30 Scales and Fails đź‘€

Hey everyone, given that there have been 30 issues so far, I wanted to do a recap article looking back at the lessons learned from the stories of all the scales and fails so far taken together. One of my favorite tweets is one by Ali over at First 1000 (I guest-posted over there a while ago):

There’s no one-sized fits all approach to building a successful startup; we’d all be doing it if there were! There are things that a lot of successful startups have done time and time again, though, and there are patterns of failure that recur in many failed ventures. These patterns of failure described by Harvard Business School professor Tom Eisenmann are core to the foundation of Scaled and Failed; there are just as many repeatable ways to kill a new company as there are to build one up. I started writing Scaled and Failed to define more of these patterns in greater detail. Let’s dive in to see what common patterns of success and failure define the scales and fails I’ve written about!

🚀 Pattern of Scale #1: Understand and Adapt Your Success

There are very few large companies that sell only one product, and most startups expand their lineup as they accelerate towards an exit. Successful expansion relies on knowing what you’re good at and why you’re good at it. What’s the reason customers come to you? How can you leverage that same value proposition in additional offerings? Nailing this when scaling up is essential.


Jagex’s landmark MMO, Old School RuneScape (OSRS), is based on an old backup of the game from 2007 and is more popular than the modern, original version. It wasn’t the easiest road for OSRS to get the internal support and resources needed to grow, but it has been an absolute hit with long-time fans. The game hasn’t been nostalgia-locked to 2007 either; the game received regular content updates and quality of life changes and a massive, dedicated community has grown around the game. The development team (usually) strongly understands the magic mix of dated aesthetic, classic mechanics, and novelty of content that makes new updates successful. They know the brand and spirit of the game inside and out even as they create new things in a 14-year-old game, keeping that old-school charm. Jagex knows what makes the game successful, and they know how to change that formula.

Jagex, which has successfully sustained a brand and game for over 20 years now, keeps the core experience and spirit the same but innovates with what they release. Content is fresh, interesting, adds to the gameplay, and most importantly is very in tune with what users want. Although their success is built on something old, it doesn’t rely on the same formula every time.


UrbanSitter was preparing to scale its babysitting marketplace since its early days, instituting a strong culture of process, structure, and documentation in the early stages of the company. OKRs and playbooks are often cumbersome when startups are looking to get their first customers, but the investment can pay dividends when the company is growing and there needs to be a way to assess a chain of actions leading to outcomes and to disperse knowledge. Repeatable processes are key to long-term success, and UrbanSitter knew exactly what it had to do to scale in new markets.

UrbanSitter’s focus on process and structure isn’t something a lot of startups do early on, but it presents enormous benefits if you’re on the path to scaling. Many startups can’t point to why they had success in one place but not the other, and that’s because they don’t know exactly what they did initially. If you can document and plan a process, there’s a chain of steps that make it easy to understand where failure occurs or success bounds from. While it may seem like a drain of resources to document and structure, it can make it easy to iterate, improve, and implement rollouts as scaling becomes a concern. It’s better to have a guide to growing on hand early than to have needed it yesterday.


SoFi started with the niche of student loans financed by alumni before becoming the neobanking powerhouse it is today. The first service was just the launching point for SoFi’s wider portfolio, starting with a variety of loans before taking on a larger suite of retail financial products. SoFi understood its market very well; there was never any question that it was the brand making modern banking for a younger crowd. It has tailored its operations around that audience, appealing to their needs across different stages of their lives. The company never lost sight of who it was serving and how to succeed with that segment.

After many years, SoFi had gone far beyond its student loan roots and created a portfolio of financial products suited to its target audience of wealthy Millennials and Gen Zs.

SoFi’s expansion across financial products reflects what many large companies do when they have the traction, brand, and capital necessary to do so. Use your initial business as a launch point to slowly diversify across an industry (and maybe across industries eventually).

🚀 Pattern of Scale #2: The Path Of Least Resistance

It’s very difficult to shift user behavior, market dynamics, legislation, and a host of other factors in our world. While it is possible, sometimes companies don’t have the capital or time to try to force these changes. It’s possible to build strong companies that equip incumbent players to innovate or alleviate current user pain points rather than create an entirely new solution.


The market fragmentation of home care for seniors and regulation of independent contractors went against Honor’s plans for matching caregivers and families, so it pivoted towards a model to empower existing players rather than compete against them. Providing back-end infrastructure and technology for entrenched home care agencies proved a lot easier than trying to take on the fragmented, relationship-based space; scaling and deployment became much easier working with partners.

Honor leveraged the technology and resources that it had developed up to that point and made a B2B play with independent agencies, which was a space filled with far fewer headaches. As they were now a partner rather than competition to the status quo, Honor would have a much easier time scaling. Everyone knows the adage “If you can’t beat them, join them!” and it rings true in this case. It won’t always be possible to unseat other companies entrenched deeply in the environment with more years of experience in an industry where technology can’t solve every problem.


Common’s decision to act as an operator of properties rather than as the owner and landlord like many other co-living startups did proved especially resilient during the pandemic and was a good strategy to derisk the idea and required less capital on their part. Margins may be lower, but it’s also a much more scalable model compared to buying up or directly leasing properties for subletting to residents and taking responsibility for payments if cash isn’t flowing in. Working with existing owners leaves fewer points of failure, and current landlords are much easier to tap into over people that might be interested in co-living communities.

While Common may have sacrificed some of the upside potential through its operator model, the risk mitigation it did through its business model has helped it flourish. Every founder will need to decide their risk tolerance in how they build their business model, but the more things that need to happen just right for a startup to succeed, the greater the likelihood of the cascading miracles pattern of failure.


A lot of transit startups have been considered the enemy of public transportation, but Via’s work in microtransit has enabled lots of cities to provide better services to their residents. Municipalities already have mountains of data and infrastructure; a public-private partnership based on technological enablement is a no-brainer when trying to scale microtransit. With enough iteration on its microtransit for consumers and the improvements this led to in its technology, Via was able to nail the pilots and proof points it needed for adoption. From there, deployment with local governments has been bridging the gap between people and public transit more effectively than Via or the cities could have done alone.

What did Via do great? While on-demand may seem less accessible to less wealthy consumers, the software platform it powers city transports systems with was a far better solution for accessibility compared to operating directly to consumers in every city, even if you have municipal backing for the service. Cities have a lot of data, a lot of existing infrastructure, and are likely already serving some people well; they’ve got a foundation, and empowering them is easier than building from scratch. Via’s eyes were on the goal of city partnerships from early on, and direct consumer service was just a step towards that.

🚀 Pattern of Scale #3: Ride The Next Wave; P-M Fit Isn’t Static

It’s difficult to know where the world is going and how it’ll look different X years into the future, but looking ahead is essential for a strong product-market fit. It’s so easy to get assumptions about this wrong, but building for right now in a transforming space can leave you in the dust by the time you’re gaining traction. If you can build for the right future, you’ll be ready to win against your existing competitors while they’re still making the shift.


Spotify did a lot of amazing things, but the most important was the vision of the future it had. Music streaming and the technology that made it the great experience it is today weren’t around during Spotify’s ideation; it helped solidify streaming’s dominance. Napster was the thing everybody missed, and piracy had driven revenue in the music industry to very sad levels. Downloading MP3 files was commonplace back then yet is just a distant memory now. Spotify was built for the future of music, and that future is now.

Spotify was forward-looking from the start, focused solely on streaming

Over time, user behaviors became more aligned with what Spotify was built for

Product-market fit can’t be decided just when a business is started, but must be assessed over time.


FlashParking was among the first companies to bring cloud management solutions to parking and has been ahead of the curve of technology there for years. It optimized for better consumer experiences from the beginning with its solutions for valet revenue management and digital payments/tipping and has since been leading digital and touchless payment in parking. With solutions that have eased the demand and supply-side experiences of parking, the startup is looking forwards to what the future of mobility may be like for parking locations: EV charging, vehicle servicing and cleaning, rideshare hubs, and more that is a far cry from today.

FlashParking’s attention to the future of the parking industry is a good lesson for anyone starting a business. Growing a startup and finding the right business model can take years, and some spaces may have a completely different landscape in that span of time. Ideally, build something that adds value now and is poised to dominate any new trends that are likely to become mainstream in the near future. If you build for just the present, you’re likely to become old news quickly in fast-moving industries. That doesn’t mean you should ignore creating value for customers right now in favor of sinking deeper into development for the future; early iterations of your product need to be useful in the present.


E-commerce has slowly been displacing physical purchasing, even in industries where the in-person experience has traditionally been valuable to consumers. Wedding dresses aren’t exactly what first comes to mind with online clothes shopping, but Anomalie was in a great position ahead of the pandemic to provide dresses when wedding boutiques and venues were closed. While wedding dress shopping in person may be a sticky custom, the customization and quality of output have built trust in the forward-looking brand as digital purchases continue to accelerate.

Even without the possibility of large wedding ceremonies with tons of guests and bridesmaids, the brides themselves likely would still be interested in dresses for their smaller ceremonies. Anomalie did just fine and continued selling to brides; it probably benefited given its digital-first infrastructure in a time when many physical retailers were struggling to adapt. In this case, Anomalie was ahead of the curve on the opportunity that the pandemic accelerated. That’s the core of a successful startup; riding a wave earlier and better than others when it comes by.

👎 Pattern of Fail #1: Good Products Don’t Always Fit The Market You Want

Products might be perfect solutions to a problem, but they may not fit the market experiencing the issue. It could be age, geography, culture, current trends, user behavior, the stage of technology today, or several other factors that invalidate a product’s fit; imagine designing a mobile-only fraud alert company with an app filled with small text and few explanations targeted at people over the age of 80. The user experience won’t be great and the solution won’t be adopted even if it has 100% success in solving the problem it wants to. It’s a hard pill to swallow, but good solutions don’t always win.


Halolife’s marketplace for death services could have had a lot of benefits, but a technology-first solution didn’t meet the needs of the market. Though behaviors are slowly changing, the people and families planning for their deaths are relatively old; the value-add is minimal given you only do this planning once so it doesn’t make sense to become accustomed to platforms you don’t normally use just to plan for your passing. Until more tech-comfortable or digital native generations are getting ready for the grave, an online solution to a complicated product will be extremely difficult.

Halolife may have made the process easier for younger generations, but it wasn’t younger generations that would be using Halolife. The problem fit the market, and the product fit the problem, but the product did not fit the market. With the referral business model and product used, Halolife was meant to have a broader reach than some of the more interesting death service startups, but it couldn’t gain any ground with its core target groups.

Death services are something that people hopefully won’t utilize very often. Even though they are overpriced and create a lot of stress, people are unlikely to invest the time to learn how to use technology well just to plan a funeral.


Before streaming came into prominence, 8tracks was king; the company let users make shareable mixes of music with their own uploadable music files. It was one of the first social music platforms, and many features from it are commonplace in other services today. The company even had an acquisition offer from Google that it turned down! Ultimately, technology and user preferences left it behind as streaming became the preferred method of music consumption with others taking the lead in that space. Society no longer wanted something geared towards their files and content but rather easy access to everything. It was a great solution for another time.

8tracks was optimized for user behavior that was prominent earlier and at the time

In terms of product-market fit, 8tracks was tailored to a type of user behavior that was disappearing: downloading actual music files.


HomeHero’s pivot attempt following tightening regulation on independent contractors was flawed given the way healthcare in the United States works. Home care has a lot of evidence supporting better health outcomes for discharged senior patients over the long term, but hospitals wouldn’t bite on a partnership with HomeHero. Since the benefits of those better outcomes are difficult to quantify and tie to a specific time period, it’s hard to demonstrate the potential cost savings to health systems of working with home care providers. The financial incentives in our system are based on more services rendered rather than value-based or outcome-driven care, so it’s hard to reconcile HomeHero’s model with the current philosophy. Even though home care does help patients and save money,  the United States healthcare system today wasn’t the right market for it.

Although value-based healthcare is slowly gaining traction, things like prescriptions and more office visits still provide the most financial incentives. Although many instances of early intervention reduce healthcare costs over the long-term, it’s difficult to predict and quantify how much is saved, making it hard to get non-medical home care in a position for growth with healthcare partners since cost reduction wouldn’t be immediately evident. As a result of the heavy focus on quantifiable outcomes and current financial incentives, non-medical home just doesn’t suit the way American healthcare operates.

đź‘Ž Pattern of Fail #2: Minimum Viable Product Needs Minimum Value

Failing fast and launching minimum viable products to test assumptions about a market are common strategies to form the foundation of a startup. They aren’t inherently bad, but people use products for a reason; that reason is usually based on the value they provide to them. If the most basic version being tested doesn’t deliver value, it can’t be successful.


While Anki envisioned social home robots serving more functions than as toys, the startup never got that far. Its releases were stepping stones in the technology used, but not in the value provided to customers. You can only sell the promise of more than a toy and potential value for so long to investors and customers, and financing did eventually fall through. Incremental innovation is powerful, but there needs to be real value delivery in the MVP.

It’s better to have a barebones MVP that delivers just a small part of the envisioned value than being fancy and complex but not executing on what it should deliver. This philosophy should be extended to products and companies overall; deliver function and do the job over anything else. That’s not to say things like design and experience should suffer in exchange for more features, as those are both critical to functionality. However, each product iteration should iterate on value for customers too, not just contribute to the development of background technology or bells and whistles that the customer doesn’t get a tangible benefit from. If it takes significant capital to finance a technological advance or design overhaul but your customers don’t see the benefit, that capital isn’t contributing to growth.


Bridj has great goals for the future of microtransit, but at its core, it only got around to offering somewhat dynamic bus routes week to week. The convenience was great, but not knowing whether your bus route would be around next time likely created a sense of anxiety and unreliability; at the end of the day, it was still just a bus and not anything revolutionary on the user’s end even if the back-end was an engineering marvel, taking in numerous data streams to plan the routes. The value for consumers wasn’t in this MVP; it wasn’t cheap like public buses nor as convenient as ridesharing, nor bridging the gap between the home and public transportation routes. Value is critical for viability.

On-demand is a huge part of that viability in microtransit and what makes a clear difference between regular bus routes and a unique solution. The lack of convenience with pop-up bus routes that change week to week would personally push me away from using a service rather than draw me in. Big data can inform the routes created, but the user experience remains just a nicer, more expensive, and perhaps somewhat better-routed bus that I can’t rely on consistently regardless of the great back-end informing the routing. With on-demand solutions and true public transit, I never worry about if I’ll get to where I’m going. This convenience and peace of mind are lost if I use something like Bridj.


Perhaps the clearest example of not delivering value is not paying out to users who had earned their money on an app. Pact incentivized people to make deals with themselves about going to the gym and used the proceeds from those who violated their “pacts” to pay out winners who stuck to them. The company was plagued by technology problems, and the core value of making money was disrupted by poor location data not logging gym time, botched execution on payments, and more. Rushing a product and not being able to provide the core value to the point where the FTC gets involved is a good way to kill a consumer startup,

Pact’s app may have been great in theory, but when the entire value proposition relies on accurate location data and timely, accurate payments, any failure of the technology is a critical failure. When individual consumers have money on the line, they won’t be very forgiving of mistakes and you could lose future potential due to poor launch conditions. Take your time to refine for consumers before it becomes a problem; the fastest launch isn’t always the best.

👎 Pattern of Fail #3: Do Things That Don’t Scale Until They Break

Doing things that don’t scale is an awesome way to build early traction and test ideas about a problem. There are plenty of examples of success when building in the early stages and it’s tied to the image of the “college kids in a garage” founders that get scrappy when they need to.

I wrote about Mid-Day Squares before, and one of the key factors in their success was how they did things that didn’t scale in the beginning to get very granular insight and build an incredible brand with their customers. Hand delivering their chocolate, taking photos in hilarious costumes to package with their bars, and spending time with customers were all great ways to win loyalty and learn what to do better. When it was time to grow, they knew what to do and moved past the actions that got them up and running; they didn’t do things that didn’t scale indefinitely.

That said, doing things that don’t scale needs to stop eventually, and it can’t be tied to the core values of a company. If you can’t scale the value you provide, the growth of a startup hits a ceiling.


The hands-on approach of matching parents and providers in Poppy’s babysitting service worked well in its test market, but the model wasn’t scalable given the costs of being high-touch and the fact that many of the community effects driving initial adoption were difficult to replicate in new markets. Getting to know the providers and Poppy helped people adopt the service, but it’s hard to get that level of personal involvement through meet-and-greets or other events in every new city. Matching also demands a much greater investment than running an open marketplace reliant on social proof; the revenues didn’t scale the same way costs did.

It’s important not only to test underlying assumptions about a problem and solution but to model the impact on these assumptions if the company grows. Can I charge the same price to customers who aren’t as acquainted with me, does the wider market have different concerns than my product addresses, does my margin get squeezed by costs that won’t be offset by revenue, what trends will affect the business on a macro scale? These are all important questions to ask and difficult to answer, but not impossible.


Luxe’s on-demand mobile valet service in cities was helpful but was an “Uber for X” startup limited by physical aspects of the parking market. There is a limited supply of parking that someone will fill, whether it be a startup or a person; there’s no incentive to offer discounts in bulk in crowded cities. Economies of scale would never kick in without the ability to affect the supply or control of parking spaces.

Many industries can’t support pricing viable for the mass-market and make money. On-demand valet parking for the average urban American is burdened by the fact that expenses rise with demand. The parking industry is consolidated and wouldn’t want to lose out on revenue, and Luxe really didn’t have any leverage with players in the space to get better pricing. If the company didn’t provide convenience and good pricing, users would just find parking themselves in lots, garages, and streets. The operators of parking spaces get their money in the end either way, so they have no reason to give any sort of discount. There was no real alternative for Luxe but to pay for the parking spaces since they couldn’t just create their own lots and garages. Users wouldn’t just stop driving into the city because they couldn’t use the app, so the unit economics were doomed.


Another on-demand startup, this one for in-person tutoring, Tutorspree failed to build in significant stickiness or switching costs to the platform and likely couldn’t meet the revenue expectations of a venture-backed startup. Other companies have made in-person tutoring enabled by technology succeed, but there were often incentives to stay on board. The ease of leaving Tutorspree was a blight on retention, and it’s hard to scale and capture new markets when you’re focused on reducing bleeding in ones you’re already in.

The idea behind Tutorspree was valid, but like many “Uber/Airbnb for X” companies it failed to build switching costs and stickiness into its platform. This was exponentially compounded by the fact it was an in-person, local service; nothing would stop good tutor-student pairs from leaving the platform and continuing offline to avoid Tutorspree’s fees. It offered some payment processing and scheduling help, but those aren’t exactly huge barriers to overcome. These barriers and tooling that offer an incentive to stay shouldn’t come later; they’ll hamstring your retention early.

At the end of the day, it’s hard to give very general lessons learned; I’m sure great companies have defied all these patterns and rules. That said, it’s easier to avoid failure and succeed in a pattern rather than as an outlier. Thanks for reading! If you found this interesting, consider sharing it with friends and subscribing if you haven’t already!