[EN] CV Android: experimental contest

With other Italian Mozilla Reps we decided to organise an experimental contest to promote Common Voice via the CV Android app. Although we knew just Android users could be able to join.

To incite users we decided to give some gadgets to the first five users in the ranking. We wrote the Rules (“Regolamento”) of the contest so everything would have been clear and in one unique place.

We decided to give two weeks as validity of this contest, specifically from 14th until 28th February (2021). Moreover, we decided to do no advertisements, but to limit this contest to the internal community (in fact, we promoted it just in our Telegram groups and bot, and on our Twitter account), this because it was an experimental, so we needed to know everything would have worked correctly before to launch the contest to the”whole Italy”(spoiler: everything worked correctly!).

We got (totally) 9 participants.

Just one of them wrote us via email, the other ones wrote on Telegram. So, we understood the most favourite way is Telegram, which is a more directly way to communicate with us.

Some users asked us to add a public live ranking, so I’ve added it on my website https://www.saveriomorelli.com/commonvoice/contests/feb-2021/, in this way users would have been able to see in real-time the own position in the ranking.

Before the contest the validated hours was 171 and the recordings hours was 212.

The first week (from 14th till 21st)

In the first days we got a considerable contributions and about 1 more hour in the Italian dataset. This demonstrate the contest, even if it’s experimental, looks to work.

Generally, in the first week, contributions have been really significant.

The better total score is about 3445 (score similar to the 2nd and 3rd users in the ranking). The 4th and the 5th users’ score, instead, is about 1873.

For other users the score is less than 100 (about 689).

In the 21st February (the last day of the first week) validated hours are 175 (more than 4 hours since the beginning of the contest).

The second (and the last) week (from 22nd until 28th)

The last week of the contest was challenging for the participants; they were keep validating until the last second useful to participate in the contest.

The top score was 10.101, following the second position: 9.572 points.

Third, fourth and fifth place number of points are very similar each other: 4069, 4021 and 3783

Sixth place scored 2369, but it doesn’t get any prize since this experimental contest stated that the prizes would have been awarded until fifth place, followed by 7th, 8th, and 9th position respectively with 2369, 851 and 31

The first-place finisher has triplicated its points just seven days before the end of the contest, from 3445 to 10.101.

At the end of the contest, the Common Voice hours validated are 181 (six more than the last week, and ten more than the start of the contest) and there was a big increase in the number of recorded hours (229 hours, 17 hours more than the beginning of the contest).

Conclusions

The contest could be considered a success, as the hours of validated and recorded records are visibly increased in just two weeks.

This graph shows how our Italian dataset has increased:

The contest made users willing to contribute to the Common Voice project have an incentive to do so since there were some prizes as a thank you for their contributions. Surely, if the duration of the contest had been longer and if also the publicity had been on a large scale since we limited ourselves to internal users of the community since it was an experimental project) the contributions, as well as the participants, would have been certainly greater.

We thank all nine participating users, both those who took home gadgets because in the top five and those who did not; all, however, have contributed to Common Voice and that’s what counts.

Ranking

Article wrote also by Damiano Gualandri