[Tails-project] Analysis from the 2018 donation campaign

Poista viesti

Vastaa
Lähettäjä: sajolida
Päiväys:  
Vastaanottaja: Public mailing list about the Tails project
Aihe: [Tails-project] Analysis from the 2018 donation campaign
Key metrics and comparison with last year:

Total raised:            74 480€         -27%
Total donations:        2239         +92%
New donors in database:        1227        +158%
Boots:                          +9%


The amount was less than last year because we had 2 huge donations in
bitcoins of 25k€ and 31k€, maybe due to the bitcoin craze of late 2017.
The very good result from this year is that we doubled the number of
donors and that more than half of these donors were new donors: that's
sustainability for the future.

The first week of the donation campaign was a big hit as always with
13.7k€ raised. But this year the number of donations between November,
December, and January didn't decrease, while last year people got tired
much quicker.

The blog posts continue to trigger a bit more donations but not that
much. This year, looking at the 7 days after a blog post, the release
notes of 3.10.1 and 3.11 were both more successful that the posts on
achievements and plans.

Note that 2017 didn't bring a lot of new features to users (3.0 rather
brought a lot of complains) while the 2018 campaign was shortly after
3.9 which brought Additional Software and VeraCrypt. From the success
metrics of Additional Software and VeraCrypt they might be useful to 14%
and 11% of our current user base respectively so I don't think that it
can account entirely for the huge gain in number of donations, nor do
the increase in the number of boots.

The fundraising team can find detailed numbers in
fundraising.git:analysis/2018/Donation campaign 2018.ods


Is it worth to accept more cryptocurrencies?
--------------------------------------------

No.

We added Bitcoin Cash, Ethereum, Ethereum Classic, and Litecoin on
December 5: https://tails.boum.org/news/new_cryptocurrencies/.

We raised only 301€ in total in these cryptocurrencies which corresponds
to 2% of the amount raised in all cryptocurrencies over the same period.


Was it worth to add a donation counter?
---------------------------------------

It doesn't seem like it was worth it.

We added a donation counter on October 28. We updated it manually from
data from CCT, RiseupLabs, and the blockchain. The update process was
time-consuming (4 hours in total) and very irregular (I sometimes had to
delay updates by more than 1 week while waiting for some data).

The impact of the counter is hard to calculate but it doesn't seem to
have limited that much the natural decrease of donations that we usually
see during the first weeks of the campaign.

I don't know how to tell whether it is a reason behind people not
getting tired between November and January.


Are people donating from Tails?
-------------------------------

Yes.

The conversion rate from /donate to /thanks (ie. PayPal donations) of
people clicking on the message that we had on /home was of 4.00% while
the conversion rate was 3.71% overall and 3.97% for people clicking on
the donate button in the sidebar.

The most motivated people where people coming from the email to previous
donors (36.79% of conversion). The opening blog post also triggered
more-than-average conversions (13.39%). I don't have data about the
following blog posts because referrers were broken during the migration
of the website to Nginx.


Proposals for next year
-----------------------

I think that this year's formula worked very well, that we should keep
it as a basis, while continue thinking about possible small incremental
improvements:

- Stick with the style that we had this year: a flashy style seems to
lead to more donations than a pretty style.
- Keep having a message on /home (last year we had none).
- Start without a counter and see how things go.
- Stick to the usual blog posts.
- Strengthen our relationship with past donors as they are the most
motivated.
- Do some usability improvements on /donate.
- Any new ideas? (with ways of checking whether they work or not)

--
sajolida