Unveiling Pix
One year and we're ready to share our amazing discovery platform with the world and what we learned building it.
What does it take to create one of the world’s largest wine platforms in just eighteen months? It takes a Herculean effort. From an extraordinary team. Made up of people willing to balance years of experience with the humility to learn an entirely new set of skills. None of us have done this before — because nobody has.
Together we’ve built Pix in the open for others to see. We've created a platform that's free to use and list for the benefit of both wine sellers and wine lovers. A platform that connects them, rather than inserting itself between them. Making Pix this way is part of our values. And so is sharing what we've done — but that's hard. Thanks to the encouragement of Erica Duecy, our Chief Content Officer, I'll be sharing our story and learning on our blog and trade content. But for now... what have we learned? What have we accomplished? What can you take away from that? And what is next for Pix?"
What we set out to accomplish
In taking on this project, I considered both the giants who came before me, as well as the existing players, who view the wine world very differently than we do at Pix. The first problem with wine is that everything is locked away in rooms where the owners see the data but use their software as a moat. Data and software are not defensible strategies. The second problem to solve (for everyone's benefit) is how we help consumers at their moment of maximum hardship — the point of purchase! And third, how do we help wine sellers connect to the consumer at that same pivotal moment. Wine sellers here means producers, who sell to sellers like importers, who sell to sellers like distributors, who sell to sellers like retailers and restaurants. We needed to solve for everyone who sells wine.
We’re building Pix in the open, and sharing what we can. Pix is not a tool whose value lies in charging you for information. Our starting point was solving the consumer’s problem; making wine discovery easier. We connect everyone, creating a meta-layer of interactions; so producers can communicate directly with consumers, partner with retailers, or collaborate with regional associations.
It’s not our job to be part of the transaction — and never will be. It’s Pix’s job to help the buyers and sellers connect. We followed the leadership of our friends at Lunchbox.io by removing ourselves from the transaction. No percentage of the sale. Zero. Zip. Zilch. The wine industry needs another intermediary like it needs a lobotomy. Pix will never sell wine and so will never compete against the companies we serve, wine sellers. We care about where the consumer is buying. It doesn’t matter whether you are a big retailer, an enterprise winery, a local wine store, or a boutique winery, our tool will help you. We will never attempt to convert customers to our market, so we can sell our own wine at higher margins. Pix needed to remove the friction between consumers and trade. There is no listing fee and no subscription, just a pure and easy search tool that helps connect buyers and sellers. It’s said analogies “can make one feel at home”. We knew we’d arrived home when we were told “Pix is Google for wine”.
What were the other big decisions we made? We decided not to replace humans, but to augment them. Instead of teaching the machine how to recommend wines, we taught it to emulate how we humans do it. That means we weren’t trying to create some kind of automated palate matcher (an absurd notion) but to learn from the consumer what they liked and to inspire them to discover new wines through content and curation. We needed to become a discovery platform. It wasn't to decompose wine into its core components and then reconstruct them again into a similar wine. It wasn't to use metadata that doesn't take into account the totality of what is in the bottle. It was to use technology to offer human answers to human problems.
The scale of the challenge is extraordinary — building a search engine for the largest consumable good category in the world is hard. As I mentioned in earlier posts, there is no central database of wine. Everyone who owns a database (except Liv-ex, who specializes in fine wine) balkanizes these pools of data as if that is a competitive advantage. Limiting access to the data creates a perpetual cycle where everyone keeps recreating the wines and their metadata over and over again. This is a huge brake on innovation and success that is holding back our industry. The moats from the incumbents have kept competitors at bay. But also keep innovation at bay that would benefit all. To give you an example, Pix features 270,000 + products. If we were to assign the metadata for those products and write content for each one, at a cost of $2-$10 per product, we would spend up to $2.7 million and we'd only be getting started as we'd be adding new wines every vintage both backward and forwards in time. That $2.7 million cost faces anyone wanting to create a tool that requires vast amounts of wine data.
This structure also means the consumer message belongs to intermediaries. We fundamentally believe that the responsibility to own the message belongs to the wine brands. It was inherent in our mission to connect buyers and sellers and allow wineries to tell their own stories, without us as the delivery tool.
But a vast ocean about wine data with a search engine on top doesn’t solve the consumer problem. It only helps people if they know exactly what they are looking to find.
It’s all about discovery
We are building the world’s first discovery platform. The word “discovery” is essential. Essentially our job is to help wine buyers discover the wine they are looking for, discover ways to buy that wine, and discover other wines to buy.
Discovery has three key elements:
It must be simple but thorough.
There are many ways of supporting wine discovery But most make it harder for the consumer, not easier. Some consumers love reviews from critics, others want consumer reviews, and some want none at all. Many of the traditional wine tools also proved to be confusing and off-putting like flavor wheels, food pairings, and flavor scales.
How we solved the problem is we met with our friends at Netflix and followed their path. Our wine team is creating thumbnails of the wine by using meta-tagging, as Netflix does. Our three keywords are style, flavor, and story.It must be useful.
We’ve created a simple search bar that helps you find things easily. Looking for wine, winery, brand, region, or variety? Type it into the search. That’s standard.
The cool magic at Pix will happen when you type in other parameters: e.g winemakers names, chillable reds, stone fruit, pizza, and more.It must be small
The paradox of choice is real. Search any brand, variety, or region and the choices are overwhelming. The key is to take a cue from other super long tail categories like music, movies, travel, and books. Again, for inspiration, we looked outside of wine to Audible, Spotify, Netflix, Amazon, Comixology, Expedia, and more. We called our friends, colleagues, and experts in these companies and the result was our Collections. These are wine groups created by Pix, as well as third parties and collaborations and they are meant to make the selections smaller and easier to discover the wines that will bring you joy. Wines with dogs on the labels. 100 pt wines from the Wine Advocate. BIPOC wineries. Great Zinfandel with under fifty case production. Etc., etc., etc.
Things we learned
We've built Pix in the open. And I've always wanted to share what we've learned so you can use our discoveries and launch your own, better startup.
First, our mantra, that we say over and over again: “Perfect is the enemy of good, good is the engine of better. Fast gets good better than good gets fast.” This phrase is fundamental to building a startup. The final products, processes, etc. are likely going to be very different from the way you first imagined it. What ensures success is good execution.
Fail fast, fail forward. Failure is inevitable in a startup. Give your team permission to fail but create a process by which you can take what you’ve learned and make things progressively better.
Usage is our oxygen. The only way to understand how well something works in software is to have active users. This helps uncover the best value exchanges, challenges, and incorrect assumptions.
The 10/100/500 rule. Creating a startup is challenging and often what we think is supposed to happen, rarely does. And facing a large task is often paralyzing. As I mentioned earlier, the key to success is execution. By breaking things down into smaller pieces, you can quickly understand the task, flaws in the process, and how to scale it. We’ve codified it into a patterned process:
The business leader does it ten times to understand how to build the framework, find flaws in their thesis, and identify patterns to solve.
Then they add limited staff to do it one hundred times and find where there were challenges they missed in their test, edge cases, and instructions.
Finally, you deploy it to a larger group of people to uncover additional hurdles. But most importantly how to streamline the processes to scale the operation.
It works with all sizes of problems and is especially helpful to create momentum.
Iteration over planning. The old Mike Tyson quote, “Everyone has a plan 'till they get punched in the mouth,” is relevant in business too. We can anticipate how things are going to work/resonate but until they go live, they are only theories. The best way to find out if something is going to work is to start small and learn and iterate. This is better than a long, protracted planning process that aims towards a target, but which will never manifest.
The light encourages growth and disinfects. This is why I like Slack and encourage a company to do as much in public as possible rather than emails and DM’s. It is often challenging and requires people to have a lot of trust and compassion. But by doing so teams can quickly collaborate on items or stave off mistakes/misunderstandings.
In the early days, hire player-coaches with intellectual humility and curiosity. Contrary to most startups that hire employees at a discount, or senior executives who bring their own teams, we believe in hiring experienced stars. Their abilities to both do and build at the same time may be more expensive initially, but the long-term cost savings and better decisions build a much stronger company. But hiring experienced leaders can only work if they have a key trait: intellectual humility. Despite their previous success, this will be the first time they have ever done this type of business, in this environment, with this new team. They need to understand that as a limitation and have a keen curiosity in learning while applying what they know.
Frameworks and processes scale, people don’t. People are the engine of a business. But they have limited time and energy. By having your leadership team create processes and frameworks, it’s easier and more effective when the business needs to scale its operations.
Teach teams to solve problems vs fixing them. Too many startups see the avalanche of problems and pull up their sleeves to fix the problems. But the avalanche never ends, it just grows and will consume your team's time until they break. But solving problems allows the teams to maintain momentum while taking on new challenges.
Wardley Maps, JTBD, Agile Software Development, and Slack. If you aren’t familiar with these business tools, Google away and spend one month on each one to understand, even in a small way, how to apply them to your business.
Challenges we faced
Remote working. We started Pix in the middle of a global pandemic — which came with advantages. The industry was awake, remote working and hiring were the norm, and we had time to focus on building. But, despite the high performance and speed of such an all-star team, building interdepartmental work and culture proved challenging. As great as remote work is, it lacks the dynamics of team building and culture. It also doesn’t allow for creative conflict and the trust that comes from resolving or brainstorming in person.
The product. Wine requires a particular balance. Having too much understanding about wine leads team members into rabbit holes. Not having enough knowledge of wine causes people to make mistakes about how to organize data, market. You want people unburdened by the shackles of meaningless wine constructs so they can innovate. But you want them to know enough so their ideas and flows make sense. And wine is so complicated and varied that you need to understand your audience, why wine works for consumers and oenophiles, etc.
Expectations. When you build a team with such talent and skill, the market expects something beyond measure. And we built the world’s best, and soon largest, a discovery platform for wine. We’ve built it at great speed. But it is by far from perfect. And that standard that people expect from us is challenging to meet. But, nevertheless, we have done something special, unique, and important.
Introducing Pix.wine (yes, we’re also still in Beta)
Go to pix.wine today and you have the world’s first wine discovery platform in all its glory — and its imperfections. Remember, this is just the beginning and it’s still a work in progress. We are only showing a small percentage of what’s in the database, as we’re honing our matching engine and adding wine sellers. We’re making our search results smarter from the data. We'll keep improving our features as well. Later this year you'll have access to our mobile app, we'll enable geo awareness, and you’ll see the engine highlight all kinds of attributes for wine searches. You’ll also see many of the world’s best experts explain regions and varieties and help wine buyers discover new wines.
You’ll see new ways to discover wines. Volcanic wines? We got you. Wines with dogs on the label? We see you. Wines made by female winemakers? Top-rated wines from Carneros under $45? When you start your journey to buy wine, please make it a habit to start with Pix.
But we’re still in beta and we want you to help us make Pix better FOR YOU. You want to add a particular wine to the system? Let us know here. If you want to see an article on The Drop, write to TheDrop@pix.wine. If you have any suggestions about how we can do better, email Paul@pix.wine. We’re still a long way from getting out of beta but we are running fast.
And finally, while we don’t have the world’s inventory into our system yet (we will), if you don’t find a wine, tell us. We’ll find it for you for free and let you know where you can buy it.
Welcome to the Pix family,
Congratulations, Paul, and your entire amazing team. I am thrilled to read all of this, and to see fruits of your labor start to bloom.