About WineLog and Super Wine Search

WineLog is a tool for keeping track of wines you've tried... to help you find more wines you will enjoy and avoid wines you won't. WineLog is a great site for wine lovers looking to rate wine, review wine, keep track their wine collections, and connect with other wine drinkers.

WineLog is also the parent company for Super Wine Search. The rest of this page will go into detail on why we decided to spin off Super Wine Search and get a bit technical about how the search works and how the two sites integrate.

Why was Super Wine Search created?

Originally search results such as those now served by Super Wine Search appeared on the WineLog "buylinks" page. The buylinks page did a good job of helping users find wines in the WineLog database. However, the WineLog database is relatively small compared to the number of total wines for sale across the internet. The question for us at WineLog then has been how to offer a better shopping experience (i.e. return more results than just those in our database) without confusing users of the wine logging tools.

Why a separate site?

Now is a good time to explain the difference between "structured wine data" and relatively unstructured "buy link data". The wines in the WineLog database are heavily structured. This means that we typically know not just the name of the wine, but also the winery/producer, region, vintage, varietal(s), tags, and lots of other information about the wine. Every piece of data has it's own place in the wine "data structure". Conversely most of the data we get from Google Base (and similar sources) comes with just a few fields: name, size, price, and a link to buy.

The structured data in WineLog is crucial to the quality of experience while using that site to manage the wine you are drinking. It allows you to be more specific about the wine in your log and allows us to do more complicated comparisons between the wine when generating recommendations. However, when shopping for wine all that is needed is enough information to confirm the wine is the one your are looking for.

Pulling the unstructured data into WineLog would have caused data quality problems that would have distracted our small team from working on other more important features. We considered putting the "shopping search" on a separate tab and using a slightly different style to indicate that the results were just for linking out to buy the wine. However, we thought it important that any "wine" showing up on WineLog be rate-able, review-able, log-able, etc. There would have been too much confusion around why some functionality was possible for some "wines" and not others.

Why not just link to Google or another site for buy links?

You can get a very similar experience by simply searching for the same terms on Google and clicking the "shopping" tab. The data is coming from the same source.

However, we believe that we can still use some of the structured wine data within WineLog and some of our own know how to provide an experience better than searching directly in Google... or using a competing site. By controlling this search engine, we can still integrate in useful ways with WineLog. Out the gate, we will show links back to WineLog when we're pretty sure the wine you are searching for is available there to log. We hope to roll out more integrations in the future.

A note on using Google Base (solely) for our data.

There are numerous companies and groups indexing shopping sites for wines for sale. At WineLog, we integrated with many of them. A year ago, we couldn't have run this site using just Google Base data. Their data wasn't that good. Lately though, their data has been as good or better than other sources. And as the largest player in the web searching space, there is little reason to think that they won't continue to be a good, reliable source for this data.

Google's size and influence also means that wineries and wine retailers should make integrating with Google (or working with services that do) a top priority.

On our end, we could integrate with multiple data providers (as we have in the past), but for now will use Google Base solely. By sticking with one data source we avoid duplication issues (a pain for us to program around and annoying for you to see) and ensure speedy up-to-date results.