The Twenty Five Doors Retro*
When I look back on how we developed and launched Twenty Five Doors, I’m not (too) disappointed. As everyone does, we made some mistakes along the way, but regardless, we followed the process. We did our research. We validated the product with paying customers. We launched the app and used the hacks to build users. We had partners.
The big mistake we made, however, was foundational. Everything else was built upon this idea; so while we followed the process and got alright results, the foundational mistake meant we never really achieved product-market fit. Ironically — and it is to the market validation process that we followed that I tip my cap — we were very right about who we thought were our customers: wine tourists. The mistake we made, however, was they were not our customer. Wine tourists were a critical customer, but the real customer was the Regional Tourism Organisations (RTOs), Destination Management Organisations, industry associations (and equivalent bodies). We knew these organisations were a critical partner, but we determined them as a channel to market; what we learned — too late — was that they were our primary customer. The foundational Twenty Five Doors mistake was wrong customer.
The impact of that one foundational mistake ricocheted through every action we took, and ultimately was the reason we failed. We had very good people; good thinkers and doers. Our first investors were incredible, in particular Ang, Scott, Matt, Cassie, Glenn, Daniel and Stan. We all worked really well, and to what we did achieve, I owe a lot to these people. I owe even more to my co-founder, Amy.
Outlined below are some of the more significant decisions we made based on the one foundation idea, and a brief commentary on how it contributed to the failure. It is written in first person — and where I use the collective ‘we’, it is for ease of reading — because even though I owe the success to the team, it was me who was ultimately leading the team, and they trusted my judgement.
Experiences — not just cellar door tastings — was something we understood early, but it wasn’t part of the genesis story. Initially it was about the map, and the concept of experiences emerged as part of the business model work we did.
But it was always our understanding of experiences that set us apart. We understood both how tourists wanted experiences delivered, AND, the challenges for wineries in delivering those experiences. This allowed us to do two significant things: (1) simplify the creation process, and (2) categorise experiences. The categorisation dovetailed into our marketing personas, such that we segmented wine drinkers based on their appetite for adventure. The experiences, likewise, were categorised according to their degree of adventure.
The Tourist Customer
Our research methodology to understand the tourist customer was exceptional. With the help of friends who specialise in survey design, we developed an extensive, multi-section questionnaire and challenged ourselves to get 200 respondents. Direct Wine Cellars (who ultimately went on to be one of our major initial investors) supported us with 2 six packs of wine as prizes, and with a solid social media campaign, in 2 weeks we had 204 respondents.
These responses were academically reviewed, hypothesises developed and then tested in one-on-one surveys sitting in cellar doors. These answers, combined with observing what happens in multiple cellar doors over a single day, gave us significant insight into why tourists visited wine regions, and what they were looking for. Not surprisingly, what they were looking for was simpler than what wineries and wine educators were attempting to deliver.
By this stage we had established 8 clear assumptions about who our tourist customer was. We then hand-selected 60 people from our network that, in our opinion and to some degree, fitted the 8 assumptions. We designed a survey around the 8 assumptions and sent it out. 56 of the 60 people responded (a good starting point), and our 8 assumptions were overwhelmingly supported.
At this point we knew why they visit and buy from wineries, how they categorised the wine they bought, and where and how the purchase took place, we knew when the bottle would be drank, how it would — or would not — form a part of the experience when it was drank. And it was based around three simple categories: (1) Mid-week quaffers, (2) Dress to Impress, and (3) Celebration. Most people initially start out with just two categories, but as they visit more cellar doors and other unique wine experiences, the middle category emerges. Overwhelming, wine drinkers prefer to buy their Dressed to Impress and Celebration wines directly from wineries they know and, even better, had visited.
And this was just the tip of the iceberg for our tourist customer research! It went on to inform everything we did, and in particular our innovative wine club idea.
The winery customer
While not to the same level as we understood the tourist customer, we did have a very detailed understanding of the winery customer. We knew this was critical for platform design, but also just general engagement and take up. If we couldn’t present a compelling case, and make everything really simple, we knew we couldn’t survive.
And to that end, our take up results speak for themselves. Across the four regions where we piloted, approximately 75% of wineries voluntarily took the effort to sign up and create a listing, and a further 66% went on to create a bookable experience. Across 130 wineries, those are results that the state and regional tourism organisations, do not even get close to if comparing engagement and listing on Australia’s Tourism Data Warehouse.
Tourism, not wine
While we understood the tourist customer intimately, we unfortunately focused too narrowly on the wine tourist niche, and it blinded me to the reality: we were in the tourism industry, not wine, and RTOs were the primary customer. This would have altered the partners we sought, our key messages, marketing campaigns, platform design, trials, and especially the onboarding process.
Our research identified three key stakeholders: tourists, wineries & industry associations. I defined tourists as our customers (they bought a membership or cellar door experiences), wineries as our suppliers (they supplied VIP cellar door experiences), and the industry association as our distribution partner and channel (they connected us to the suppliers and thus received a commission).
And it worked. To a degree. It worked because we took the time to deeply research tourists and wineries in a structured process. We definitely understood our customer and our supplier. But that didn’t extend to our channel partner. And not understanding our channel partner meant not understanding the competing forces and challenges of the tourism industry.
And those competing forces were massive!! Tourism is funded by a patchwork model combining Federal, State and Local government funding, ad-hoc grants, sales commissions, membership and advertising. Overlaying the patchwork funding is a large, international ‘system of systems’ (the Online Travel Agencies — OTAs — and all the interconnecting infrastructure, the Australian Tourism Data Warehouse — ATDW — included), that directs how tourism activities (ie VIP cellar door experiences) are packaged, promoted and sold.
(If you are interested, read my environment scan on the use of technology to support regional tourism activities, and in particular how the current infrastructure supports wayfinding).
So, because I studied the wrong customer, I failed to equally understand the real problems facing our channel partner and real customer. And not understanding their problems meant I spent our time fixing issues in the supplier & customer channels, often at times making things more difficult for the team. I made decisions that made us swim against the currents of the massive, international tourism system.
Currents I would have understood if I realised that we were in the tourism industry, not the wine industry.
One of the best ideas that emerged from our research was how to innovate the traditional winery wine club model. And, even better, it was the perfect marketing strategy to keep tourist members — who only visit wine regions once every 12–36 months — engaged after they left the region.
It was simple: 25 wineries creating a limited edition, monthly wine club offer at (their) member-only prices, with free freight and no obligation to buy, exclusive for Twenty Five Doors members. It was like being a member of 25 winery wine clubs, without any obligation to buy. We also worked with one of our early supporters — Peter Bourne — who provided members with his Keen Eye Pick — basically, the monthly wine pack that stood out above the others, that someone with a keen eye would snap up. It did work: there was a correlation in the data between Peter’s recommendation and sales.
I personally loved the brand, and received a lot of positive comments on its evocative imagery and the intrigue of what was behind 25 doors. But not everyone loved it, and I had many people insist I needed to change it.
What I learned from this is that brand, quite literally, can be anything — so long as you get the emotion of the imagery, language and messaging right. At times we got it right, at times we didn’t. But how far we got — selling to the wrong customer! — I think is a testament to what we did do with the brand. It evoked the core discovery element of the brand, and the road as the primary imagery was a stroke of genius: everyone loves wine regions because they love the sense of adventure and the road trip they are going on. We just captured that essence (and by we, I really mean Matt Jones & his crew at his creative agency!).
There was a heap more behind the name and brand, but none of it really matters. Ultimately I saw how the brand has to be infused with real values, otherwise, it is useless. And sometimes, not everyone is going to love the brand.
The importance of team at any stage of development can never be underestimated, but at pre-start up and pilot launch, it is critical. These people create momentum that helps delivery.
Our friends and family investor team was fantastic. In particular Scott, Ang, Matt, Cassie, Stan, Glenn & Daniel. They all put in the sweat equity, but we made it fun. It was a great adventure we were all going on together. We had weekly workshop sessions where tasks were divided and we conquered! I would not start anything again without this level of commitment.
Wine club, not wine experiences in cellar doors
My first fundraising pitch confused the market. The idea was a ‘wine club that aggregated real cellar door wine club offers’, and a portion of the sale funded the development of the cellar door marketplace (app). A member — wine lover — of the Twenty Five Doors wine club was supporting real wineries with real cellar doors by buying their wine through our wine club. As an added bonus, they were also funding a new app that would bring more people to the cellar door to learn through experiences, not words on a screen, why wine is so incredible.
The theory appealed to my values and sat well with our for-purpose approach to the business. Everyone won. But it was convoluted. Nevertheless, I started building the wine club by talking to wineries about the wine club AND the tourism platform (app), but as we are all busy, the conversation quickly became just a conversation about a wine club.
And so when we launched the app we had to spend a lot of time explaining to wineries and all stakeholders who we were. Even once they used the platform and had the “Wow!” factor of the map, they were still confused.
Failing to hear
This one hurts a lot. Not because I ignored the message, but because I deflected the message, thus missing its true gold. And, unfortunately, the people around me justified my response. We all missed the gold.
I was part of a regional accelerator for the Hunter Valley, the pitch competition prize was to pitch in Sydney against a Manly accelerator. A small, but good crowd, an interesting set of judges, all set against the ocean backdrop of Manly beach.
I delivered my pitch well. That was never the problem, but afterward, in Q&A, things went odd. One of the judges kept asking me who our customer was. I replied saying it was the tourist. She challenged me, I kept my calm and clarified that is was the tourist through the industry association. She challenged again, but by this point, she had already bought in her association with a similar failed ‘wine club’ business that her partner had started, and I was increasingly uncomfortable with how things were unfolding, so I deflected all further challenges and invited her to take the conversation up off stage.
Things ended pretty smoothly, and despite her challenges I genuinely wanted to have the follow up conversation. I didn’t understand where she was coming from. As we left the stage I had two close mentors and a senior Hunter Valley / Newcastle Angel Investor approach me. All of them applauded how I handled the challenges, and all of them told me not to take too much notice of her comments; we all agreed they were driven from her previous bad experience in a similar situation and she had good intentions.
In hindsight, we were wrong, she was right. I don’t know if she was right about the specifics of our business, but she was right about the customer question. We didn’t know, with 100% laser focus, who our customer was.
Not sufficiently aggressive / proactive
In an industry like wine and tourism, ‘hustle’ is a dangerous concept. I remember listening to a podcast where they discussed the characteristics of founders. I took particular heart to this because the investor described a ‘different horses for different courses’ theory, highlighting how a founder like Travis Kalanick (Uber) was necessary to break down doors in building a ride-share market in big cities. She went on to say that Travis would have been terrible in many other industries. As someone who struggles with the ‘hustle’ of a start up, I liked this idea because it meant I could be me. And our results show that wineries and industry associations definitely appreciated my style. I learned later the stats for the State Tourism Organisations equivalent efforts to achieve a similar outcome: we outscored them 6–1.
But, in hindsight, one of the errors was not being more aggressive; and by aggressive, I mean proactive. I was always worried about the load on myself — and it’s flow on impacts to my family — that I kept things at a level that was always manageable; but that wasn’t enough irons. I needed a few more to keep momentum going.
At this stage, had I properly understood the customer, I could have expanded quite easily the number of regions we were talking to, and found more traction in each region. I might have benefited by triangulating the influential stakeholders — all of whom were beholden to a broken funding model — and uniting them in partnership through the app — an app for their region — that solved three of their problems: revenue, wayfinding and data.
I made many decisions about capital that were wrong. The biggest was not properly consulting my co-founder how we would fund the business once the capital we raised ran out.
I assumed, and acted in isolation to secure, that an offer of a loan from my family, combined with some of our money, was the best way forward. Once we were properly established and money was flowing, I reasoned, we would return the loans and see the returns by way of a predictable salary.
Nothing was wrong, in principle, with this plan: the mistake I made was not getting agreement on the plan. It put pressure on the family situation and future decision making.
I wish I could say which were the Red Herrings and which were not, but I seriously have no idea. But I know I spent time chasing Red Herrings. It is something I have learned about myself through this process; I am very good at creating possibilities, but turning those possibilities from Red Herrings to Golden Gooses often eludes me.
The wrong customer meant we researched, built and tested for the wrong onboarding process. We focused entirely on the tourist and winery, and not on the region. I even made the mistake of not segmenting regions in our initial master report which slowed down our ability to report region-level data.
Understanding RTOs as our customer, I would have definitely included them in our design and testing process. Ironically, building as we were for customer/supplier/channel, we got it pretty right. Our ‘Wow!’ factor was the map, but thinking as I was, I sold the wrong message to the wrong customer, and built an onboarding process that was cumbersome for staff in regional coordinating offices.
When researching the requirements definition we were clearly focused on the winery needs. As the supplier, we wanted to make things very easy for them. But listening — and agreeing — to everything they said was a mistake. When we were consulting, the concept of bookings was still in its infancy in Australian cellar doors, and accordingly wineries said they needed 48 hours notice to provide VIP experiences. It sounded reasonable, so we hard-coded the requirement. This was probably my only major ‘visitor-as-customer’ error; unfortunately that one error had massive implications for our revenue model.
The error was significant because tourists were using our key feature — the map — as part of their in-region planning and wayfinding tools, but with the average stay less than 48 hours, and the majority of visitors only commencing their research once in the region, the hard coded requirement for the supplier became a shot in the foot to our revenue model. The reason you become a member is to get VIP access. To get VIP access, wineries need 48 hours notice. Tourist leaves in less than 48 hours, and hence just use our platform as a free planning tool.
Thinking through how this error may have been different if we had correctly identified our customer, I think we may have got timely input from RTOs / industry associations about wineries and customers. But then again, perhaps not. One thing we definitely learned about RTOs and industry associations as we kept developing the product was how little real-time data they had. They all had hunches, trends and visitation surveys, but no one had (and no one still has), the real-time, live data to know what is actually happening.
To our credit, we identified this mistake quickly, but just struggled with fundraising and other issues to get this fixed in the code. We developed a perfectly good alternative in partnership with a few wineries, but we never got the chance to deploy it.
Assuming we could come back to UX
I read something, somewhere, that quoted Paul Graham (Y Combinator) and/or Ried Hoffman (LinkedIn) as saying: “If you’re not embarrassed by your first product, you’ve spent too long on it.”. I also took this to heart with our MVP. Functionally, it was fantastic — still better than most destination management software in the market— and the big buttons and intuitive UX made it work well in demonstrations and with winery users, however, the UI was terrible and there were some significant, overall UX factors that we had to just accept.
We did accept them, and they weren’t the things that made a massive difference, but toward the end, when the MVP was nearly 4 years old, we did struggle against newer apps. Ironically, had we included a designer up front and not assumed we could come back to UX, we could have made decisions that would not have cost any money, if anything, they might have saved us money.
Scoping & phasing
This wasn’t a major failure, but in the early days of build, we didn’t understand the best way to scope and phase what we were doing, which, I am sure added significantly to the build cost and time (a lot of which was done in kind).
What this did do, however, was give me a good understanding of why Agile development is so important, in particular the Back Log. I used this to go on and develop our own Agile approach.
Wine club, not wine experiences
We launched the wine club 8 months before we launched the app that connected members to the wine experiences. We used a wine event at Bagrangaroo (Sydney)and despite the event being a bit of a debacle (unrelated to us!), we had good success; a conversion rate of about 13% of people who visited our stand.
But these numbers mislead me. It validated that tourists were willing to pay, but also blinded me further to who the real customer was. Nevertheless, we were able to use those tourist conversions in our early conversations with wine regions. At that time, I could see that a central issue was going to be the quantity of people we could bring to the table; what I didn’t understand was that I was in a quantity, not quality, conversation because I set the ‘tourists’ as the frame of reference.
With this frame of reference, the number of members we had in the wine club became the only conversation. The general regional tourism business model works using destination marketing campaigns, locally marketed events and partnerships (like ours) to attract people to visit the region. Without big tourist numbers, we were always just bit-players.
So further reinforcing the fundraising confusion I created, the wine club and focus on wine club members also confused wine regions.
The correct framing of the conversation needed to be about partnering with them to improve the capacity of wineries to deliver quality cellar door experiences and using the power of technology to connect the right tourist with those experiences. Ironically, had we been having that conversation, we would have been perfectly positioned to ride the wave create by Wine Australia and the “$50m Package”. As it turned out, we were always just outside the core focus.
Understanding that the customer was the industry association / RTO, the conversation would have been about their problems (capacity building and wayfinding — or connecting visitors to experiences), which opened the door to the real conversation about destination management that we needed to be having. With the destination management conversation, our strengths — digital wayfinding and transformation — were their weaknesses. And positioning ourselves as were, as a long-term partner, our value proposition would have been to grow the region’s success together.
So our launch was to the wrong customer, with the wrong angle.
Understanding where our customers were when they would make the decision to engage with us: GFWS v In Region
We launched the MVP at the Good Food and Wine Show in Melbourne and Sydney. It wasn’t a cheap launch — and in hindsight, wasn’t worth the relative money spent compared to what was left in the kitty — but it went well. The team came up with some novel ideas — a Wine Walk — which wineries and visitors to the Good Food & Wine Show enjoyed. We didn’t get the same level of membership conversions as the pre-launch (I think it was about 8%), but our brand was being recognised, we had more validation of paying customers (this time for something new), and we had lots of interest from wineries.
But again, here is an example of planning and executing a good campaign for the wrong customer. We did alright, but I should have done something similar, but at a tourism trade show. We would have had a different message, with a different app onboarding process and a different set key features. Ironically the app wouldn’t have been that different, just we would have sold the features differently.
What the Good Food & Wine Show demonstrated to me was that we had to shorten the decision timeline for members. At a show they can’t necessarily imagine when they are going to visit a cellar door next; when they are in the region, they know exactly when they are visiting a cellar door. Today or tomorrow. Unfortunately, although I eventually came to this understanding, the focus on the visitor-as-customer meant it took longer than it should have.
Post Pilot Launch
No fundraising strategy
The plan post pilot launch was get more regions on board and try to attract the attention of angel investors. Again, our numbers weren’t bad (definitely not great), but we used them to sell the vision.
In hindsight, this is where things became wobbly and I didn’t follow a process so methodically. It was around the time our second baby was due and I was facing a number of challenges. The ‘onboard-and-flap-my-arms’ strategy was never going to deliver the investor we needed, but I was still driven by the visitor-as-customer frame and too focused on how to build visitor members that I couldn’t see the wood for the trees.
I attempted to bring structure to fundraising when I engaged a broker. This was a good decision, but the mistake I made was believing that fundraising would then take care of itself. No. I needed to do more, and unfortunately, I gave him the wrong message, for the wrong customer, to sell.
He and I worked hard to find the message, and we got close, but by that time we were deep in the 2nd Melbourne Lockdown and COVID had scared away any interest I had in a now, 3-year old start up in the tourism industry.
Not sufficiently selling or promoting the pivot
A key early learning from the pilot — and my first clue as to who the real customer was — was watching stakeholders respond to the map. We proved during the pilot we could get wineries to quickly onboard, and we realised that if we pivoted to a deeper partnership with the region, including making some critical onboarding changes, we could quite easily provide something very new and of tangible value for their websites: a digital map and personalised itinerary creator. It was embedded in any website as a button with a URL to their secure and private part of the Twenty Five Doors engine. Or put another way, we could be their Destination Management and region app.
This is the pivot, with more foresight about the real customer, combined with a fundraising strategy, could have bought us the capital we desperately needed, but unfortunately we continued to talk to the wrong investors about the wrong customer, and hence the wrong problem. We were still looking at wine industry related investors, not travel/tourism investors. More effort selling the pivot to the right investors would have revealed my biggest oversight: not integrating with the Australian Tourism Data Warehouse (ATDW). This decision, which flowed from looking at the wrong customer, was the other foundational failure.
Too many eggs in too few baskets
Not integrating with ATDW was not a decision made in absence of our data. The problem was, our data set — tainted already by wrong customer, was further muddied by working with 3 Victorian regions and 1 South Australian region. Although ATDW is a big thing in SA, it was still early days and regions weren’t as integrated then as they are now; in Victoria, businesses have to pay to list on ATDW, and hence no regions or wineries were even talking about it. We followed our Victorian biased data and went the other direction because what we had — functionally and in our product roadmap — was better.
Wrong customer. Wrong data set. Wrong decision.
Again showing how decisions compound, had I been more proactive in working with regions, had a more coherent fundraising strategy and had more irons in more fires, I would have had a greater data set to inform decision making.
None of the mistakes outlined here were ill-made decisions or rash actions. They were planned and based on the available data we had at the time. I remember when I was a graduate at the Department of Defence, straight out of university, I spent time with a General that defined leadership as the courage to make decisions based on the available information, and also to say they made a mistake if new information became available that changed the assumptions underpinning the first decision.
That idea has always resonated with me. Perhaps knowing this idea, I was happy acknowledging my incorrect decisions as the new data became available, but I was too slow in responding. Yes there were extenuating circumstances and lots of good possibilities that justified the decision to continue with the MVP v1.5 plan (a half-step toward the full pivot to regions as customer), but ultimately, I was too slow executing.
The obvious lesson of this experience is about customers. I am now obsessed with understanding the customer experience and, more importantly the system it sits within. This can be most easily seen through journey mapping of each stakeholder and a systems analysis of interactions between each stakeholder and their journeys. I went far in this exercise to build Twenty Five Doors, but not far enough. I had an inkling, but I didn’t follow it with the same obsession I did for tourists and wineries. Big lesson.
The other lesson is definitely about people. Specifically talking to people. Without experience in either the wine or tourism industries I often felt like an outsider and was challenged by my engagements with wineries and regions; this did limit the information I learned from them. I learned a lot, but had I been able to bridge the ‘outsider’ gap I felt when talking with them, I would have heard more.
So maybe that is actually the lesson. Hearing. Not just listening. Hearing. Being brave to clarify what doesn’t make sense and hearing the real conversation, that may, or may not, be going on. Because, in hindsight, I had plenty of evidence to support what I now know.
*If this was a real retro, the entire team would have been involved. As it is, we still might do a team-wide retro, but for now, this is my personal founder & CEO retro. Furthermore, it doesn’t celebrate / look at the wins we had. I allude to them, but decided to focus on lessons learned.