Many of us have been curious how overall traffic to the website has changed recently. So, using old URLs and piecing together various bits of information from the website, I created a graph illustrating the number of ratings on Ratebeer by month since 2010.
It seems like 2019 sees us a little below the 2014 monthly average, after which things flatlined, until a slight uptick in 2018.
Note: I had to delete one month (Nov. 2017) because it showed a million ratings, which doesn’t seem trustworthy given how much of an outlier that’d be compared to other months (~10 times the monthly average prior to it).
Method: the way I did this was to use the old rating page, find a beer I rated on the 1st of each month, and click the “Delete Rating” button which then brings you to the page that tells you the rating number from the whole site that that beer constitutes. Thanks to @fiulijn for the tip off on how to navigate this.
I’m actually wondering about the reasons for the flat performance in 2014/2015.
I would say that 2016 is somehow when we started seeing many trending breweries releasing 1-2 new beers every week, making it much easier in many areas to tick more beers, but the growth doesn’t seem to be steeper than in the past.
Ah, this is great, I’ve wanted to do this for years.
Out of curiosity, I overlaid general craft beer production on top of this and it’s quite close, at least until 2018. (There is no 2019 data yet but I assume it doesn’t have this precipitous drop.) data
The numbers do not impress at all-.
There is members that I know about that has been on this site since 2002.
See your graph start on 2009, thats ok, I assume before that the numbers where much lowe than 40K/month.
If you say that now there is an average of 10000 users (that is low compared to what it shoud have been) and if you say further that there is an average of 80000 ratings a month.
That gives an average of 8 ratings pr. user/month.
I do not have the number of actual users, but I do hope it is higher than 10K after 17+ years.
Interesting stuff surprised 2018 was that high in relation to past years, geography of raters may just have shifted why some of us feel we don’t see many beers rated in our areas.
Would be interesting to see unique raters by month of this timeframe. So folks who rated at least 1 brew to know if greater access to beers means the supers raters can now rate 300 beers a month versus 100 or if actually had more people doing the occasional rate.
You can fudge this by using Ratebeer’s numerical URL system. That is, you can take a beer URL like this and change that number at the end and get a new beer. This number increases in-order, so you can make a timeline of beer adds. Here are a few data points:
Beer 100000: Added Nov 2012
Beer 500000: April 2017
Beer 700000: Feb 2019
Beer 761000: Yesterday (Aug 22)
It shouldn’t be too hard to write something to crawl the major milestones and build a basic graph, but I’ll leave that to someone else. There will be a little bit of annoyance in crawling the paginating system to get to the oldest rating, which is loosely indicative of add date.
I did that “crawler” last year.
It was searching by increments of 50,000 (admittedly easier for me because the beer-edit page has the date).
It was impressive to see, even with the “limits” of RB penetration, how fast the craft (?) beer world is growing.
It took us 8 years to put together the first 100k, now we need 8 months for 100k.
It also shows how easier is for young punks like mcberko to rake ticks compared to when Per had to fly to inner Asia to find a new tick
But I was skipping RB and mentioning UT because in the last years we can’t assume anymore that the numbers are representative.
In the past it was a safe bet to say that, say, 90% of the beers released were added to the database.
But now that percentage may be a variable, for example it could have been 60% in 2016, 50% in 2017, 40% in 2018.
So, knowing the number of new beers on RB may (very likely) not be a good indicator anymore.
My million ratings stat shows that there’s a reasonably long tail. The guys down at the 1000-level are still putting some effort in. Everyone that “retires” is being replaced by someone who is still active, eventually. (I.e. “death of the site” would be a complete stagnation of that stat, which isn’t happening".)
You have to remember that long tails are long. Even if you need to go through 50000 accounts in order to find just 2%, or 1000, active raters there is that number available. When I did a quick probe of random account ids, it was clear that an enormous proportion were inactive. An enormous proportion weren’t ever even used for any rating even once (like >90% or something stupid). But yeah, your number looks like a good first stab at a guess.
Let’s assume that 20:80 active/inactive ratio which exists at ranges 1700-2000 keeps on. In that case to triple the 839 active users you would have to take 10390 users. That goes beyond the top list we have access to, but since user no 7500 has 100 rates we can assume that final couple of thousand users have below 100 rates each.
So I’d say even that scenario seems optimistic. I’m pretty sure this site has less than 2000 active contributing users. And that’s what it feels like. It’s not like we’ve seen influx of fresh blood here.