One of the biggest challenges in building a product is prioritization.
When you have 1,000 things bombarding you and demanding your attention, it’s hard to figure out what is most important. You have your customers telling you one thing (vocal minority), your team and investors telling you another (selection bias), and your data telling you another (confirmation bias).
Everyone is an armchair critic.
The result is that I see a lot of products that do way too many things poorly. Many times this is the result of multiple undeliberate pivots in a short period of time. Don’t get me wrong, I’m pro-pivot — I think the general idea encourages fast iteration. However, I think sometimes entrepreneurs pivot poorly. They are so stuck on the product they’ve built (sunk cost fallacy) that they fail to extricate themselves from the trenches of building a product to reevaluate the overall problem they are trying to solve.
They end up with a Frankenstein of a product that does nothing well, a jack of all trades, master of nothing. The product “pivot” ends up being a hammer in search of a nail, rather than the other way around. Because they’ve spent so much time building out the product, the mental bias is that there has to be some value to the product they’ve built; it is easy to mistakenly correlate time spent building the product with market need.
The entrepreneur’s job is to synthesize all of this feedback and test your assumptions in a methodical and deliberate manner.
Without a product True North, you end up wasting time optimizing on a local maximum and building a Frankenstein Product.
(Stay tuned for a follow up post on a framework to avoid building a Frankenstein product.)
One of the prevalent trends in startups recently is the bundling of products into one, with a combined value greater than the sum of its parts.
This is not a new concept; product bundling is a strategy that has historically been effective in selling products and maximizing economic value.
Product bundling is most effective when bundling high volume, high margin products, commoditizing the individual products and increasing the value of the bundle as a whole. This means that bundling is particularly effective with information and digital products.
In some cases, bundling of inferior products can actually be more compelling than individual, unbundled superior products.
Here are 6 examples of bundling and the value bundling adds:
Example 1: Pokémon Cards
Value: Randomness and Potential Expected Value
This is where I expose my nerd card. Pokémon cards were a big fad when I was a kid. We would buy booster packs of Pokémon cards, which included a random unknown selection of 10 cards.
In the case of Pokémon cards, the randomness of the unknown selection of cards within the booster pack drove large sales of the booster packs. My friends and I would buy 10 booster packs and hope to get 1 holographic Charizard card. As a business, it would be difficult to sell the more common cards individually, so by bundling them with the potential of the rare card, you increase the value of the booster pack and drive huge sales of the booster packs. The booster pack’s value was greater than the sum of its parts.
The value of the “rare” cards was driven by scarcity of and demand for the rare physical product rather than any real value itself. And the potential (read: expected value) of that booster pack was driven up by the fact that those rare cards were bundled with the common cards.
Example 2: Subscription Services: Birchbox, Quarterly, Craft Coffee, Trunk Club, Foodzie, etc.
Value: Discovery; Convenience; Curation
Subscriptions services bundle individual products into a recurring box of unknown products. The idea is that you receive products that you wouldn’t have otherwise known about or bought for the following reasons:
Discovery: You find products that you wouldn’t have discovered otherwise.
Convenience: You can try out the products without having to leave the comfort of your home. Then, the hope is that you will buy the full sized product (conversion to purchase) or become a loyal customer (customer acquisition to lifetime value).
Curation: You get advice for which products are better, or a better fit for you. This is especially powerful when products are commoditized and the market is saturated to the point where it is difficult to choose which one is better.
Example 3: Information: NYTimes, Twitter, News.me, Summify, Social Weekend
Another type of bundling is the bundling of information and content.
One of the reasons people read the NYTimes is because they trust the NYT editors and writers to give them the perfect balance of information and news, both the must-know breaking news as well as the interesting reports on stuff off the beaten path.
I personally get all of my news through Twitter. Others might get that information through News.me as a filter on top of Twitter. We all choose our own filter bubbles and arbiters of information. We trust those arbiters to bundle information together. In some cases, we value it so much that we even pay for it in the form of a subscription.
Example 4: Products/Interest Graph: Pinterest, The Fancy
Startups like Pinterest and The Fancy let people create digital bundles of products and images. The analogy that Pinterest uses is the digital scrapbook. The value proposition here is that I am interested in what products people with similar tastes like.
The startup that can successfully convert this bundling to purchasing behavior will be well-positioned to win. The Fancy is getting good early traction around converting curated digital bundles into single purchases.
Ultimately, I think the opportunity is to convert curated digital bundles into buying whole bundles of products.
Example 5: Media: Music (CDs, Napster, BitTorrent, iTunes, Spotify/Rdio)
There were two major shifts of bundling in the music industry.
The first was the shift from the physical to digital medium, from CDs to MP3s. The critical effect here was the idea that consumers could suddenly get songs a la carte, and buy individual songs separate from the bundled album. There was more value in debundling the songs from the CDs, because the consumer didn’t get much value out of the predetermined album bundle. They wanted to create their own mixtapes, burning MP3s on CD-R’s.
The second major shift was from MP3s to streaming services like Spotify and Rdio. Here, the idea of owning any form of content is completely removed. People now listen to playlists of songs they don’t own, curated by other people they many times don’t now in real life.
Example 6: Financial Products: CDOs
Value: Hedging Risk
Finally, an example in the financial services space. Not to oversimplify the financial crisis in 2008, but collateralized debt obligations (CDOs) gained instant notoriety in 2008 for being the financial instrument that caused the financial crisis. CDOs were originally intended to hedge risk; specifically, mortgage-backed securities of varying risks of default were bundled together into a larger financial product.
Ultimately, however, rather than hedge risk, CDOs incorrectly masked risk. The models behind the CDOs did a bad job of predicting how risky the underlying securities were, and the ratings agencies did a bad job of realizing this.
Conclusion: Bundling and the Commoditization of Complements
As a parting thought, the overarching idea of bundling is the idea of the commoditization of complements. By bundling products, you commoditize the individual products. This in turn increases the value of the bundled product you are selling.
The classic example is the hot dog stand. Take a hypothetical situation with two competing hot dog stands. One of them charges for the hot dogs, ketchup, and soda individually. The other hot dog stand bundles them all together into a “happy meal.” The value of the bundled product is much higher if the customer would have bought all three items anyway.
In fact, bundling may actually be a good antidote against “race-to-the-bottom” pricing, where the only competitive advantage among commoditized products is price.
Bundling is an incredibly effective tool to increase the value to your product offering.
The luxury of scale is a double-edged sword. While scale can bring huge network effects you can leverage for fast user growth of new products, it can also mislead a big company to go after the wrong opportunities and paralyze a small company.
So when developing a new product, make sure you build a compelling first time user experience (FTUE) that does not rely on the benefits of network effects. The FTUE is important to get a push out of the gate, but iteration and designed-in network effects are what will win the game in the long run.
When Big Companies Assume the Luxury of Scale
Big companies that have achieved scale and take scale for granted are many times trapped into slow moving and myopic decisions that optimize on a local maximum.1 This is a topic that’s covered extensively in The Innovator’s Dilemma.
Big company that did this poorly: Google
Google Wave and Google Buzz both failed because they assumed the luxury of scale. They assumed that the large user base of Gmail and Google Search would transfer directly to a social product. Google mysteriously assumed that, just because it was a Google product, people would automatically sign up and use it regularly. It made no effort at integrating with email or any particular effort to ensure engagement other than sticking “Google” in front of the product name.
The FTUE was bad because Google neglected the core user behavior for both of those products: both are inherently controlled private experiences. Yes, Gmail is “social” in the sense that you are emailing other people, but they are private in the sense that the user expects data generated by Gmail to be private.
Big company that did this well: Zynga
Zynga was able to turn effectively a non-scalable game studio business into a massively scalable company with huge network effects and metrics-driven best practices. They understood the core value proposition and the first time user experience to make sure the user is hooked with each new game. Every single time Zynga launches a new game, they know:
And most importantly, the folks at Zynga understood that you have to design to scale by designing in network effects. If they had designed assuming scale, they would have failed much like Google Buzz and Wave did.
When Small Companies Assume the Luxury of Scale
Small companies that have not achieved scale, yet assume massive distribution without understanding the core value proposition of the product, often end up with a product that has a bad first time user experience (FTUE) and fails to convert users into highly engaged users.
I’m a big believer in the KISS school of design. Test single hypotheses and control for every other variable to isolate a cause-and-effect relationship.
Small company that did this well: Foursquare
Foursquare’s first iteration of their app was solely focused on getting people to check in to a venue. No other bells and whistles in the feature set. They tested that core use case / value proposition and found that it is a hugely viral action that taps into our need to humblebrag (more on this in a separate blog post).
They assumed no scale or network effects. I still remember the key complaint about Foursquare when they first launched — that you need a certain number of friends for the product to be really interesting. I think the threshold number of users that Facebook found that users became engaged users was around 15 friends. So Foursquare used Twitter’s existing scale and pushed a lot of actions to Twitter by default. Also, Foursquare’s badges and mayorships features helped improve the single-player experience.
Small company that did this poorly: Hot Potato v1
In the first version of our Hot Potato app,2 we fell into the trap of assuming scale for the FTUE. As a result, we failed to pinpoint the main reason why a product feature is compelling from the first time the user loads the app. The single-player experience was bad. When users landed inside an event conversation stream, they saw an empty feed the majority of the time. The app worked well for large scale events like WWDC and SXSW, but failed for the long tail of event feeds on the app.
So when developing an app, make sure the FTUE is compelling. Users have limited attention spans. Without a core focus, everything looks like the most important thing to focus on. The edge cases begin to look like the core focus.
Make the product useful for your first users and understand its reliance on network effects before designing for scale.3
If you chase two rabbits, both will escape.
Hot Potato was an app we built for people to have conversations with like-minded people around live events. ↩
One of the more fascinating patterns in online communities is the emergence of highly engaged communities and behavioral patterns built on old, and many times, bad technology.
The classic example is Craigslist. Despite the old and sub-optimal design, people continue to use Craigslist and have created behaviors of their own to meet their needs. Check out this excellent diagram by Andrew Parker that shows all of the startups that have been built around each section of Craiglist.
The existence of these highly engaged communities and the fact that the crowd coalesces around a common cause despite technological hurdles is a strong indicator for the market need.
Also fascinating are the patterns of behaviors that emerge within those communities. For example, Style Forum is a highly engaged community of people passionate about mens’ fashion. Within this community, some behavioral patterns have emerged:
Only recently have the forum administrators added new features based on that user feedback, but the new features are big hits. There have also been startups launched to build very focused products around those specific behaviors.
This means that you should have just the right amount of structure in your product to define the core action, but have enough slack to enable serendipity.
An example of this is Twitter. Early Twitter users created the @ reply, retweet, and #hashtag.1 When the Twitter product team recognized these behavioral patterns, they productized these behaviors and made them part of the core product to own that behavior.
Another example of this is what my friend Joe did at The Fancy. His product did one thing very well (posting pictures of objects), and then he closely observed what his users wanted. He found that his users were scouring the internet to find a place to buy the product and posting the link in the comments.
So not only did he build a product that captured that serendipitous behavior, he recognized the behavioral patterns and productized around that. The result is the new e-commerce offering you see on pages like this one.
Maximize structure around your single core behavior and minimize structure around edge cases to maximize serendipity.