I had some fun with the statistics for social genealogy sites in Average Size is a Statistic. I used the public numbers for Geni and We’re Related to deliberately produce some contradictory results; the average size is less than 10 and more than four million. Size is a statistic, and you know what they say about statistics…
The comprehending reader has already grasped the real message; not only is
average size a statistic, but average size
is a metric that does not fit social genealogy sites.
The ultimate social genealogy site would contain everyone; it would have some
six billion users, with six billion profiles for the living and let’s say 54
billion profiles for their ancestors. The profile / user ratio would be 10 and
the average size
would be sixty billion.
Some interesting metrics for social genealogy sites are number of users,
number of profiles, the resulting Profiles / User (P/U) ratio, the number of fragments, and the
size distribution of these fragments.
Quality of the data, and the ability of
the system to process it all in timely measure are interesting metrics.
Most
social genealogy vendor show little interest in either data quality or
application performance - and thus hardly attract the users that are. These
sites are a way for users to connect through genealogy, and millions of users
are happy to do so, however slow the application or doubtful a distant
connection.
This article does as most vendors do, and largely ignores both application performance and data quality, but that does not mean that these issues are unimportant. As the social genealogy market matures, users are likely to attach more importance to application features, performance and data quality.
Number of users is obviously an important metric. There’s nothing social about a social genealogy application without its users. More users is better, and vendors eagerly publish how many million users they have.
Number of profiles is another number that vendors are eager to boast about. The larger this number is, the more likely you are to come and check it out, to see whether any of all that data relates to your research.
The ratio between the number of profiles and the number of users is an
interesting figure. Vendors that boast about the millions of profiles and users,
implicitly publish their P/U ratio.
As How Geni beats We’re Related pointed out already, that ratio
is a numeric indication of how usable and enjoyable the application is. That
GEDCOM supports increases the P/U ratio is only right; as it makes data entry a
lot easier.
Vendors are already publishing the number of profiles and the number of users, but merging of trees into larger fragments ensures that there are less fragments than users. For the published data to be really useful, vendors should publish the number of fragments too, so that we can calculate the profiles / fragment and users / fragment ratios.
The profiles per fragment (P/F) ratio is the social genealogy alternative to
the average genealogy size. It tells you how large the fragments are on average.
Larger is better.
The Size is a Statistic article used the P/U ratio to calculate
average tree
size, but the P/U ratio does not tell you how large the fragments
are, it merely establishes a lower bound. You need the P/F ratio instead.
The users per fragment (U/F) ratio is a measure of how successful the site
has been at connecting users - or getting users to invite their family.
The U/F ratio is an indicator of how likely you are to connect to other users, as
well as to how many users you will connect when you do.
Many social genealogy sites use family relationships for viral effect; users are encouraged to invite family members, and they in turn are encouraged to invite their family members.
Vendors regularly publish press releases boasting about their numbers, they pull stunts involving the American president, and set up FaceBook groups to create web buzz around their products and services.
Social genealogy sites are web 2.0 applications, and all the usual metrics, such as number of active users, time spent on the site, the page interaction quotient, and user retention apply.
Vendors are likely to regard many of these numbers as trade secrets, but retention figure can be estimated from the number of visitors and the
number of unique visitors published by sites such as Quantcast.
For social
genealogy sites, the P/U ratio is both a rough indicator of user retention and
and a more meaningful metric anyway.
Social genealogy sites are subject to the network effect; the more people have joined a site, the more valuable it becomes. As it becomes a more valuable resource, more users join the site. They add their data, making the site a yet more valuable resource, and so on. The effect boils down to new users choosing what early users have chosen already. This is a positive feedback loop that ensures that the largest sites grow fastest, that the most popular application become yet more popular.
The network effect rewards early entrants and makes it hard for newcomers to capture even a small share of the market. Hard, but not impossible, as there is a lot vendors can do to affect their own metrics positively.
Social Genealogy Success lists some ideas on how vendors can improve their social genealogy metrics.
Copyright © Tamura Jones. All Rights reserved.