01438 741177         thewinesociety.com

Sharing cellars on Cellartracker - WIKI



@peterm, Some of the wine I scanned did come up with an alternative as you’ve suggested and I don’t think I have a lot of obscure wine either .Possibly a few but not many! I store wine in 3 different places. A wine fridge which isn’t nearly big enough, under the stairs… decent size, Victorian house, and in the garage. I began to input a few weeks ago what was stored in the wine fridge (where I keep the lay it down, good stuff). 50% of that wine didn’t scan straight off with approx. 30% of that 50% being recognised as a “Burgundy” etc… the others not recognised. I then tackled the understairs and more were scanning successfully, however not all. Then I hit the garage last night! Literally every second bottle I picked up there was an issue with again just wine I have either bought from a producer themselves, a small wine shop in France or Italy or from a trade supplier. It got pretty frustrating to be fair so I didn’t finish it. I did obviously change vintage on the ones that needed changing but less than half of the non scanning wines were recognised as anything at all. So yes I stick by my 40% average. Thanks for your suggestion on how to log new ones though… sometime I may get around to it.


Sounds like the issue you are having is with the label scanning facility, not the actual CT database.
I’ve rarely had an issue finding a wine on CT when typing into the search facility.


Yes I log in the Wine details to search the data base. I’ve tried the photo system which is based on Vivino but it’s limited.


@Russ that’s what I’ve been doing using the Vivino software :weary:…!



Vivino relies on people uploading photos, which is fine for supermarket wines which seems to be the demographic of the users. Certainly I find it great to check out the latest Waitrose stuff. For more “upmarket”(?) wines there doesn’t seem to be the information, but it’s there, on CT, if you type in. I guess the CT users have more age worthy wines. I don’t think the photos on the CT are picked up by Vivino, although CT use they’re app.
Most of this is second guessing on my experience and I’m a non techy middle aged man, so the bits and bytes are lost on me.


PS the guys at Ct are very helpful,. I’d drop them a line for tips.


adding my personal thoughts too

It would be unfair to criticise the CellarTracker database for wines not being matched via the Vivino label recognition software

In my experience the CellarTracker database is very comprehensive and clean, while the Vivino one was quite the opposite. I know they have put in a LOT of effort to clean that database (imagine how many duplicates will be generated by millions of consumers adding wines and creating new ones when they don’t understand how to read labels) but there will still be issues.

While the label scanning app is a fun and useful extra for CellarTracker, nothing beats searching manually for the wine. If you add just a few details in the ‘add wine’ box you are almost certain to find what you want.


Also I have found that using the barcode scanner in the Cellar Tracker app works pretty reliably.



Sorry, I don’t know anything about Vivino…

Based on my own experiences of manually adding wines I don’t believe 40% of your wines are not already in the CT database, however I can accept that the Vivino app doesn’t find them.

Have you seen

I can also recomment contacting Eric LeVine, creator of Cellartracker. He is very responsive and wants to fix any problems users have; use the contact form on the site or post on the CT support forum. If the app is supposed to work with CT and it’s not then the CT will want to know about it in order to fix it.

Best of luck!


LiveForLunch here kids.


I’ve found the barcode reader quite variable - sometimes it pulls nothing, and sometimes completely the wrong wine! However like others I find that if you type it in you find what you want nearly all the time.

Of course to improve the database what you then need to do is scan the barcode of the wine you just found manually - but I’m as guilty of not doing that as anyone.


To be fair the only wine I did not find recently by typing the name was the Kamocsay Ezerjo, which was then quite easy to create based on another wine. Except for the database does not seem to recognise the Ezerjo grape, so I might have taken a shortcut…


That is a very good wine at an amazing price! CT is pretty US focused - so I’m not surprised Hungarian wines and varietals aren’t well covered. That’s the beauty of it though - us Europeans can build on the database and hopefully broaden the exposure of all the users to the delicious wines available outside of the established global exporters.


The barcodes on wine labels cannot be relied on as the same barcode is used on different wines/sizes/vintages from the same producer and different barcodes are used in different territories. See http://support.cellartracker.com/article/10-about-upc-and-ean-barcodes

If you are using CTs own barcode option (where you produce a label with CTs own barcode to stick on your bottle, to scan out consumptions) that should be 100% accurate.

However, I understood the previous poster was using an app that identified a wine from its label image, and teh same wine can have different labels in different markets.


Post on the CT website forum Wine Data Errors & Corrections and they’ll add Ezerjo to the database


I’d like to weigh in a little here as someone who has spent a little TOO long already looking at wine data.

Classifying an individual wine in a database is INCREDIBLY difficult as there is no accepted single source of data such as the ISBN for books. It means that every single winery, importer, retailer and service-provider will have their own database and their own system, and matching them up can get even more tricky.

In terms of its comprehensive nature and its level of detail, there is virtually nothing to compare with CellarTracker (IMHO) but it relies not on automation, but on user input to keep that way, so it is also fallible.

I gather that they have partnered with Vivino who have managed to stand out from the wine-app crowd to get MILLIONS of installations and tens of millions of scans of bottles in their system, so they have a truly market-leading solution to identifying wines from their label images. However, their data is also sourced from users, and the average Vivino user is not as well versed in wine as the average CellarTracker user, and so many errors can creep in to their system, and staying on top of that to keep data acurate (‘clean’) is a mammoth task. It would also probably be true that the average Vivino scanner drinks a different set of wines to the average CellarTracker cellar owner, so there could be a substantial portion of wines for which no existing label image will be recognised.

In this case, there are many reasons that using the label scanner will fail:

  • the image matching fails (often relating to the lighting quality of the image taken in a cellar)
  • the image matches the wrong wine even though the correct wine exists, but the image doesn’t take you there because the correct wine has no associated image
  • the image matches the wrong wine even though the correct wine exists, but the image doesn’t take you there because the matching engine has been incorrectly trained
  • the image matches the correct wine but that wine doesn’t exist in the database yet

It is great that the two can now work together but we cannot assume they will be infallible - as CellarTracker points out at the end of the article shared by @peterm:

Remember, label recognition is like auto-correct: it’s a labor-saving device, but sometimes you have to assist it. If you get stuck, just reach out and we’ll be happy to help!

On top of this we now have the issue of Barcodes. As Peter also points out, these also vary by producer, country and retailer. We like to assume they are unique, but in practice there are many reasons they are not. In addition, while all producers add some kind of label, only those destined for multiple grocers or specialist retailers using SOP systems will bother to add barcodes. This makes that set of wines biased towards the mass-market end.

Maintaining a cellar record is not an easy task, and in many cases is really not worth it unless you keep things for long enough that you need to keep an eye on drinking dates, or you keep wines for their value and want to monitor that.

Keeping and sharing this data can be fun, but it is also a lot of work, and while CellarTracker, Vivino and others make this easier, it still involves a reasonable amount of elbow grease.

All this is to say that I really appreciate all the work you have done to catalogue your wines, and also that you support not only this community with that work, but the overall global communities for those sites too. If it is a lot of work, it would help if you were willing to share that back to their communities too by helping with the matching of images and barcodes to wines, or creating the missing wines and vintages … but that really is something for the true wine geek.

I hope this helps to put all this in context and to say say how amazing that despite all this, we have records of our cellars to share :slight_smile:


I’ve found the barcode scanner a little hit or miss in the past so I don’t use it. Manually entering details in CT isn’t too onerous as I’ve only had to do it a couple of times for over 500 bottles.


Really interesting thanks Robert - what’s your view on the more expensive options like Cellar Watch? I have friends with rather deeper pockets than mine who swear by it - but they are managing very large (multi-thousand) bottle collections of high value wines.


I don’t have a view on those, to be honest, as I have never used them.

There is no shortage of systems for managing cellars … as long as the cellar you are maintaining focuses on ‘fine wine’. These wines are those that are traded not just drunk, and because they retain and gain value, there are margins to be made selling services to the owners.

CellarTracker is not exclusively high-end and so although the concept of a ‘cellar’ still tends towards those with ‘vin de garde’ it happily includes plenty of other wines in the system. Vivino is even more that way, designed for wines that are ‘popular’ and therefore it is more about the shopping experience than the value/trading experience. Hence the poor overlap between all these databases.

I know that I am not really interested in the higher end as I collect no wine other than for delayed gratification, and I suspect many here will feel the same. The Society is focused on bringing drinking pleasure to members, not financial speculation :slight_smile:

However, if you deal a lot with en-primeur, particularly if you are a collector, then you may get value back from those systems.


Yes that’s why I like CellarTracker - there are people with wine worth more than a house on there which is fun to ogle - but most users are much more modest. I’m with you all the way on buying to enjoy - I do buy quite a bit EP but that is simply for me to build up a cellar for me to experience over the next few decades.