[metrics-team] Tor Browser initial download numbers

[Tommy Collison]

On 2/9/18 12:27 AM, Karsten Loesing wrote:
> Hi Tommy,

Hey there,

>
>> I just spent some times with application download data in Tor Metrics.
>> I'm talking to a funder and it'd be great to be able to say "Tor Browser
>> was download x times in 2016 and y times in 2017 -- and this z percent
>> increase is proof that we're good at being a nonprofit."
> Just in case you did not yet have the numbers, here's what I came up with:
>
> 32,730,720 downloads in 2016
> 32,842,844 downloads in 2017
>
> That's a relatively small increase. But still, I'd say that having a
> relatively stable number of downloads is a good sign.

Yep, those are basically the numbers I came up with. Being able to point
to a steady number is great.

>> Can you give me an insight into how reliable the download data is? How
>> are partial downloads treated? (You may have seen my conversation in
>> #tor-project just now.)
> I'd say these numbers are pretty reliable.
>
> Regarding partial downloads, we do count an aborted download as 1, but
> we're not counting resumed downloads. I'd say we're slightly
> overcounting, but not much. In any case, we're doing the same thing in
> 2016 and 2017, so the absolute numbers are comparable.
>
> Hope this helps.

Okay, that makes a lot of sense, thanks so much for clarifying!

Tommy