<s>
Website	O
popularity	O
is	O
commonly	O
determined	O
using	O
the	O
number	O
of	O
unique	B-Protocol
users	I-Protocol
,	O
and	O
the	O
metric	O
is	O
often	O
quoted	O
to	O
potential	O
advertisers	O
or	O
investors	O
.	O
</s>
<s>
A	O
website	O
's	O
number	O
of	O
unique	B-Protocol
users	I-Protocol
is	O
usually	O
measured	O
over	O
a	O
standard	O
period	O
of	O
time	O
,	O
typically	O
a	O
month	O
.	O
</s>
<s>
This	O
is	O
because	O
an	O
increasing	O
number	O
of	O
people	O
now	O
access	O
websites	O
through	O
multiple	O
physical	O
devices	O
with	O
different	O
IP	B-Protocol
addresses	I-Protocol
,	O
and	O
thus	O
may	O
be	O
counted	O
multiple	O
times	O
.	O
</s>
<s>
Websites	O
that	O
track	O
users	O
using	O
cookies	B-Application
instead	O
of	O
IP	B-Protocol
addresses	I-Protocol
are	O
not	O
safe	O
either	O
,	O
as	O
some	O
users	O
regularly	O
delete	O
cookies	B-Application
from	O
their	O
devices	O
,	O
and	O
others	O
may	O
use	O
multiple	O
web	O
browsers	O
to	O
access	O
a	O
single	O
website	O
.	O
</s>
<s>
This	O
means	O
that	O
for	O
a	O
typical	O
news	O
site	O
,	O
for	O
example	O
,	O
which	O
people	O
might	O
visit	O
more	O
than	O
once	O
a	O
day	O
to	O
keep	O
up	O
with	O
breaking	O
news	O
,	O
the	O
reported	O
number	O
of	O
unique	B-Protocol
users	I-Protocol
may	O
overstate	O
the	O
number	O
of	O
different	O
people	O
by	O
a	O
factor	O
of	O
four	O
.	O
</s>
<s>
Similar	O
to	O
the	O
TURF	O
(	O
Total	O
Unduplicated	O
Reach	O
and	O
Frequency	O
)	O
metric	O
often	O
used	O
in	O
television	O
,	O
radio	O
and	O
newspaper	O
analyses	O
,	O
Unique	B-Protocol
Users	I-Protocol
is	O
a	O
measure	O
of	O
the	O
distribution	O
of	O
content	O
to	O
a	O
number	O
of	O
distinct	O
consumers	O
.	O
</s>
<s>
A	O
common	O
mistake	O
in	O
using	O
Unique	B-Protocol
User	I-Protocol
numbers	O
is	O
adding	O
up	O
Unique	B-Protocol
User	I-Protocol
numbers	O
across	O
dimensions	O
.	O
</s>
<s>
A	O
Unique	B-Protocol
User	I-Protocol
metric	O
is	O
only	O
valid	O
for	O
its	O
given	O
set	O
of	O
dimensions	O
e.g.	O
</s>
<s>
For	O
example	O
,	O
a	O
website	O
may	O
have	O
100	O
unique	B-Protocol
users	I-Protocol
on	O
each	O
day	O
(	O
day	O
being	O
the	O
dimension	O
)	O
of	O
a	O
particular	O
week	O
.	O
</s>
<s>
With	O
only	O
this	O
data	O
,	O
one	O
cannot	O
extrapolate	O
the	O
number	O
of	O
weekly	O
Unique	B-Protocol
Users	I-Protocol
(	O
only	O
that	O
the	O
Unique	B-Protocol
User	I-Protocol
count	O
for	O
the	O
week	O
is	O
between	O
100	O
and	O
700	O
)	O
.	O
</s>
<s>
However	O
,	O
website	O
administrators	O
who	O
can	O
track	O
unique	B-Protocol
user	I-Protocol
traffic	O
over	O
a	O
longer	O
period	O
of	O
time	O
can	O
build	O
up	O
a	O
reliable	O
view	O
on	O
their	O
performance	O
against	O
direct	O
competitors	O
within	O
the	O
sector	O
.	O
</s>
<s>
Online	O
businesses	O
tend	O
to	O
have	O
a	O
static	O
conversion	O
rate	O
between	O
unique	B-Protocol
users	I-Protocol
and	O
new	O
business	O
clients	O
.	O
</s>
<s>
When	O
calculating	O
movement	O
of	O
unique	B-Protocol
users	I-Protocol
through	O
the	O
conversion	B-Protocol
funnel	I-Protocol
,	O
the	O
same	O
time	O
period	O
must	O
be	O
used	O
at	O
every	O
step	O
.	O
</s>
<s>
A	O
visitor	O
is	O
sometimes	O
referred	O
to	O
as	O
a	O
unique	O
visitor	O
or	O
a	O
unique	B-Protocol
user	I-Protocol
to	O
clearly	O
convey	O
the	O
idea	O
that	O
each	O
visitor	O
is	O
only	O
counted	O
once	O
.	O
</s>
<s>
The	O
purpose	O
of	O
tracking	B-Application
unique	I-Application
visitors	I-Application
is	O
to	O
help	O
marketers	O
understand	O
website	O
user	O
behavior	O
.	O
</s>
<s>
The	O
measurement	O
of	O
users	O
or	O
visitors	O
requires	O
a	O
standard	O
time	O
period	O
and	O
can	O
be	O
distorted	O
by	O
automatic	O
activity	O
(	O
such	O
as	O
bots	B-Application
)	O
that	O
classify	O
web	O
content	O
.	O
</s>
<s>
Estimation	O
of	O
visitors	O
,	O
visits	O
,	O
and	O
other	O
traffic	O
statistics	O
are	O
usually	O
filtered	O
to	O
remove	O
this	O
type	O
of	O
activity	O
by	O
eliminating	O
known	O
IP	B-Protocol
addresses	I-Protocol
for	O
bots	B-Application
,	O
by	O
requiring	O
registration	O
or	O
cookies	B-Application
,	O
or	O
by	O
using	O
panel	O
data	O
.	O
</s>
