Google analytics tells me that over 2% of my visitors are using Chrome 1 or 1.5, is that accurate?

Question

This puts both Chrome 1 and 1.5 in the top ten Chrome-using visitors to my site, and they are surrounded purely by at least double-digit versions, and mostly v32.

It also gets users of these versions above my self-imposed "1% of total" cutoff for browser support. Just barely. (Actually, it puts it at just over 2%.)

Looking at a broader range (the past year), 1.5 doesn't even show up in the top 20, but 1.0 is #12 (at .6%, well below the cutoff).

Am I to take these stats to heart and consider testing my upcoming (JQuery/JQM/Bootstrap 3 riddled) site with the first version of Chrome? I didn't even think that was possible.

Or is this some weird Analytics anomaly?

Azimuth

that sounds suspiciously like bot traffic to me. – Sundew

I thought that too, but do bots go around in beat up old browsers like that? – Azimuth

no, bots don't usually run in browsers. They are usually standalone scripts that fake things like user agent – Sundew

I should add that this is a fairly busy site, with 73,000 visits last month, so I'd think cracking the top ten would be no mean feat. But I just don't know. – Azimuth

That's kind of what I was trying to say... that the bot was "riding" in the digital equivalent of a beat up 78 T-bird. Maybe undercover. So in other words, this could either be old bots that started crawling when 1.5 was a legit browser, or just faked a user agent that happened to be very old...? It's just an awful lot of bot traffic, I would think. – Azimuth

Well, I just wanted to be clear, that a bot is (usually) a standalone script, independent of the browser. The User Agent is just a string it would specify in the request header. It can be any value you want. It can say "Bob's brows-o-matron 2.0" for all it cares. But I get what you're saying.. it could be an old bot written back when that was the browser version to fake. More likely, it's a more recent bot that's just faking old browsers. In either case, – Sundew

IMO it's a lot more believable than an awful lot of people using a really old browser.. another alternative is perhaps it is a legitimate bug in the GA interface, perhaps incorrectly parsing some new browser version that came out. If this were the case though, you'll likely hear about it in GA's issue logger. I personally haven't heard of this from others though. – Sundew

If it's a bot the visits should have more in common (OS, Screen resolution, clustered around a few geo locations etc) so you might want to check for those kind of patterns, too. – Dunford

Okay, I checked with a few sites I monitor and I can confirm that there are tens of thousands of visitors with Chrome 1.0 and 1.5 roaming the internet (lots with ecommerce transactions, too, so they are unlikely to be robots) . I assume those were Chrome versions without auto-update feature. So yours might actually be legit users. – Dunford

I suppose Chrome 1 is newer than a lot of v1 browsers out there, so it's not outlandish to think of these as possibly legit. If so, sadly, it's just enough users for me to be concerned enough to test it. Unfortunately, I don't know of a way to do so currently... I've been using Browserstack recently, which only goes as low as v14, and even Google's OldApps site doesn't offer the download anymore. Anyone know of a way to test this version? – Azimuth

Were these visits from chrome mobile rather than chrome desktop? – Tremblay

Answer 1

I noticed this also... I have about 1.0% on GA showing Chrome 1.5 and 100% (according to GA) is "Mobile" and 97% is "Samsung" brand. Taking a look at this page - http://www.webapps-online.com/online-tools/user-agent-strings/dv/brand125499/samsung-galaxy-note-iii - makes me wonder if GA is counting strings like Version/1.5 Chrome/28.0.1500.94 as Chrome 1.5.

My best guess would be Google Analytics just screwed up the parsing of the newer Samsung user agent strings, which (again, according to that webpage) looks to be Chrome 28. At least that's what I'm going to mark it down as.

(Sorry for the 4 month late response. Figured other people might be Googling this also.)

UPDATE: These are the entries for "Mobile Device Info" on my GA results for Chrome 1.5...

Samsung SCH-I545 Galaxy S IV
Samsung SCH-I337 Galaxy S IV
Samsung GT-I9500 Galaxy S IV
Samsung SGH-M919 Galaxy S IV
Samsung SM-N900V Galaxy Note 3
Samsung SM-N900A Galaxy Note 3
Samsung SM-N900P Galaxy Note 3
Samsung SM-N900T Galaxy Note 3
Samsung SGH-I537 Galaxy S4 Active
Samsung SM-P600 Galaxy Note 10.1 2014
Samsung SM-T520 Galaxy Tab Pro 10.1
Samsung SGH-I337M Galaxy S IV
Samsung SM-P900 Galaxy Note Pro
Samsung GT-I9505 Galaxy S IV
Samsung SCH-M919 Galaxy S IV
Samsung SM-T900 Galaxy Tab Pro 12.2
Samsung SCH-R970 Galaxy S IV
Samsung SCH-R970X Galaxy S IV
Samsung SM-N9005 Galaxy Note 3
Samsung SM-N900W8 Galaxy Note 3
Samsung SM-T320 Galaxy Tab Pro 8.4
Samsung SPH-L720T Galaxy S4
Samsung SM-P905V Galaxy Note Pro

All of these show potential user agents with Version/1.5 Chrome/28.0.1500.94, which should be interpretted as Chrome 28, not 1.5.

Celiaceliac

I am agree. I got some weird version (A lot of from 4.0, and some from 1.5 and 1.6) on my analytic as well...They are all mobile and tablet. Most of them are Galaxy as well. I also researched that, Chrome 1.0 was release in 2008, and never release any subversion like 1.5 and 1.6 at all...so I am pretty sure there is some problem in the analytic... – Durrell

I'm still surprised Google hasn't updated that yet. I mean it took them about 1.5 months to figure out IE11's messed up useragent, but this one has been screwed up for what, over 7 months now. And it's a Google browser. I think Analytics is Google's red-headed stepchild. – Celiaceliac

@philtune I'd like to add that versions 1.6 and 4.0 represents 3% and 1.95% respectively of my Chrome traffic. Are those supposed to be poorly interpreted UAs as well? – Misanthropy

@JoãoCunha, I can only assume. I just ignore them on mine because they account for <0.5%, and most Chrome usage is usually the latest version within two weeks of its release. But in your case 3% might be pretty significant depending on your total traffic, so it's worth Googling around (or make some custom Analytics filters) to make sure you know what group that is representing. – Celiaceliac