The Fine Wine Data Problem: Past, Present and Future

The Past

Currently the fine wine market is pretty rough compared to what it once was a few years ago, but it is improving slowly. Anyone who’s viewed the Liv-Ex 50, 100 and 1000 indices recently can attest. The fine wine market was booming up until late 2022, interest rates were low, so high net-worth individuals looked for other, alternative assets that could grow their money. Fine wine became a highly attractive target, which led to massive amounts of new investment capital to flood into the market, which in part drove the prices up.

The COVID-19 lockdowns did a lot for the wine market, wealthy consumers had their fancy holidays cancelled and couldn’t visit the high-end restaurants (the horror), this led to far more money being spent on “staying in”. Home wine cellar kits surged in popularity and fine wine companies were forced to modernise, buying trading and auctioning all had to be brought online overnight. This all created the perfect storm, people had far more time to research and treat wine like a properly traded asset.

During the wine as an asset boom, it wasn’t just individual buyers, it birthed a whole new industry of unregulated wine investment firms. Companies promising high returns, often comparing the Liv-Ex 1000 to the S&P500, it all looked great on paper. These companies would take your money, then build, store and manage lucrative fine wine portfolios.

Unfortunately nothing lasts forever, starting in early 2023, interest rates were raised again and investors started realising that they could get their classic, risk free, traditional asset based income again. Fine wine doesn’t pay dividends or interest, so investors tried pulling their money out. People found out the hard way that wine is no where near as liquid as traditional stocks, their wine was either over-valued or simply unsellable. This pushed the whole market into a downward spiral.

I’ve checked the profit of some of the leading fine wine merchants, from Berry Bros to Armit Wines, to Seckford, of the 17 on my short list, 9 were profitable in their last financial year, but the market is out of the hole, and slowly going back up, strengthening naturally, rather than into another boom.

The Present

Let’s fast forward to now, this where I join in on the fun. Let me start off by saying I’m not a wine drinker, but I find fine wine and the fine wine market very interesting for a variety of reasons. History, culture, wine fraud (just thought I would add that I’m interested in reading about wine fraud, not committing it myself), just to name a few. I’m sure some would say trading wine without ever drinking it is sad, but in my eyes, it’s an advantage.

Originally my plan was to get into wine collecting, buying wine in bond that I believe would go up in price if cellared for a few years, but after some proper market research, I found a much better option for me personally, a potential gap in the market, an easy gap to fill, or so I thought.

Anyone who has spent any amount of time using Wine-Searcher to look at fine wine has seen the same thing, large price mismatches, the same case of Bordeaux being sold for £500 by one merchant and £700 for another. Now this could be for a variety of reasons, a broking client or merchant trying to quickly liquidate stock, wines leaving their ideal drinking window, pricing errors or just plain competition. There are hundreds of wine merchants in the UK alone, when all prices are aggregated in one public place, merchants and users often start undercutting wars to steal the eventual sale. Unfortunately it’s not feasible to continue using Wine-Searcher to find these gaps manually, this is why I chose to build a better way to scan and track these prices over the greater bonded market.

Unfortunately, the market isn’t especially liquid, (I’ve been trading for a few months now and it’s better than I thought, but still not great), which isn’t ideal for me a recent university graduate trying to find a respectable side hustle I can talk about with regular people without them looking at me strangely, I also sell digital goods online, but that’s a topic for another time. This is where spotting active buyer demand comes into play, tracking where people are actually willing to pay more than the market average, at the most basic level, finding where the buyers are waiting.

So all someone needs to do is create a system that takes merchant data and compares it against the current active demand, simple right? Nope, wrong, it’s taken me months of work to get up to this point, I’m still finding issues, but also finding solutions too. The first big issue was getting the data, currently I track over 90 merchants, pretty much all of them store and export data in different ways, so custom data pipelines had to be written for every merchant. To make things worse, there’s barely any standardisation in the wine world, “Ch. Léoville Barton 2016”, “Leoville-Barton 2016 12x75”, and “Leoville Barton, St Julien 2016” are all the same wine. I will do a separate write-up all about my entity resolution struggles here at some point, as this is already very long, but this has been one of my biggest issues. What I currently have is pretty accurate, but it took me a very long time to get to this point and I’m still trying to improve it as we speak.

The Future

Now although it would be very convenient, I can’t see into the future, so I don’t know what’s going to happen, so I’m going to use this section to talk about how trading has been going so far. I’ve been working on this project for so long, I wanted it to be perfect before I actually started, but eventually I realised that that wasn’t possible. I believed that the only way I could continue to iron out the bugs was to start using it for real, I committed to buying my first case and just as expected, I found a number of issues I’d never thought of before, but, I learned from them and made changes.

I bought my first case in May, since then I’ve bought 15 cases in total, it’s not all been smooth sailing, whilst working on the project previously I had tracked 14 different demand indicators over multiple different platforms, this allowed me to see how liquid the market was, 90% of these buyer signals stayed in place for over a month, this gave me confidence that this idea would work. But a couple of the cases I’ve bought so far had their interest fade during transit, ironically, including my first one.

I was faced with a dilemma, I need the market to be both liquid to let me sell cases easier and not liquid to allow me to have the time to physically move the stock. Most of the sites I sell on require the case to be in their custody before I can sell on their website, which is fine, my quickest turnaround has been 11 days. But on transfers that take their time, especially when I’m going from one storage provider to a completely different one, LCB to Octavian for example, these take so much time that the initial buyer interest could fall through before it arrives, this then means I have to list the wine and then it must sell naturally. This is when I need the market to be liquid, this has happened a couple of times now, it did make me worry slightly, but thankfully these have sold eventually.

Whilst working on the project I was worried that there wouldn’t be enough viable opportunities to make this project worth it, when I start running it each day I realised that this could be worth it, then beyond that more recently I had a dry patch, where I had no opportunities over a certain profit percentage for a while, but then after a few days it got better again. The wine market is in a funny place at the moment, but I think if you’re smart, you can still do well in the space.