Mixing it up: Developing and implementing a tagging system for a content-rich website which uses aggregated content from multiple sources.

And re w Le wis S e nior We b C onte nt M anage r, Victoria and Alb e rt M u s e u m
Linked In h ttp :/ b it.ly/ O p AN F
/ h

M ixing It U p :
D e ve lop ing and im p le m e nting an in-h ou s e
tagging s ys te m for a conte nt-rich we b s ite

M u s e u m s and th e We b
7 Ap ril 201 1

And re w Le wis

 S e nior We b C onte nt M anage r at th e
Victoria and Alb e rt M u s e u m Lond on, s ince
2008

 R e s p ons ib le for te am m anaging conte nt
on we b s ite

 Inform ation s cie nce training with p u b lic
lib rary e le ctronic s e rvice s b ackgrou nd

Th is s e s s ion is h op ing to give you …

 A gre ate r u nd e rs tand ing of wh at tagging
can d o and h ow th at m igh t b e u s e fu l for
you .

F orm at of th is s e s s ion

 Backgrou nd - wh at we we re trying to d o,
and h ow we ap p roach e d it

 An au d ie nce e xe rcis e in tagging...

Th is s e s s ion d is cu s s e s

 Brie f b ackgrou nd to V&A we b s ite
re d e ve lop m e nt
 Atte m p ts to cros s -conne ct vis itors with
conte nt th e y m igh t not re alis e we h ave
 D e ve lop ing a tagging s ys te m for s im p le
s u b j ct clas s ification
e
 Th e h u m an is s u e s arou nd tagging conte nt

Backgrou nd - re d e ve lop ing th e V&A we b s ite

 R ationale
– To im p rove s ite vis u ally (e as y b it)
– To m ove to organis ation b y u s e r’s inte re s ts
not b y inte rnal d e p artm e ntal s tru ctu re
– To u p d ate s ite to b e re s p ons ive to typ ical
googling/ ocial-re com m e nd ation b e h aviou r of
s
we b u s e rs
– Move toward s op e n s ou rce and inte gration of
u s e r-ge ne rate d conte nt and we b s ocial m e d ia

H ie rarch ie s

8
b e ta.vam .ac.u k

H ie rarch ie s

h e lp

10
b e ta.vam .ac.u k

H ow d o vis itors land on ou r we b s ite ?

S o u rce : G o ogle An alytics fo r V&A s ite : 1 Janu ary – 31 D e ce m b e r 201 0

C lu s te ring arou nd u s e rs ’ s u b j cts of inte re s t
e

H ow d o vis itors land on ou r we b s ite ?

Mixing it up: Developing and implementing a tagging system for a content-rich website which uses aggregated content from multiple sources.

3 le ve ls of control

 E ditorial – ab ility to s p e cify an article m u s t
ap p e ar u nd e r a s u b j ct clu s te r
e

 S emi-automatic – ab ility to d e fine th e ke y
s u b j ct of th e article or s u b j ct clu s te r as th e
e e
b as is for clu s te ring conte nt

 A utomatic – s e arch logic analys e s th e te rm s in
th e conte nt and u s e s algorith m s to m atch it with
oth e r conte nt

Au tom atic and
s e m i-au tom atic

Au tom atic and
s e m i-au tom atic
E d itorially s e le cte d conte nt

Au tom atic and
s e m i-au tom atic

6 tagging fie ld s ...
§ “ pers on” to re fle ct th at vis itors wou ld b e looking for inform ation ab ou t
s p e cific artis ts , d e s igne rs , h is torical and fictional figu re s

§ “ place” to re fle ct vis itors inte re s t in art and d e s ign as s ociate d with
p articu lar locations

§ “ technique” to s u p p ort vis itors trying to u nd e rs tand artis tic and d e s ign
p roce s s e s

§ “ period/s tyle” to re fle ct an inte re s t in artis tic m ove m e nts or s ignificant
nam e d h is torical p e riod s

§ “ date” to re fle ct th at vis itors m ay b e inte re s te d in work p rod u ce d at a
p articu lar p oint in h is tory

§ “ free text” – th is was inclu d e d to allow fle xib ility and b e ab le to e m p loy
te rm s not cove re d by th e m ore s p e cific cate gorie s

A lot of s p re ad s h e e ts ...

23

Th e p roce s s of tagging conte nt

– Au d it re ve ale d ove r 4,000 p age s on th e m ain s ite

– Tagge d ind ivid u ally b y re ad ing th e p age and
d e cid ing th e m ain s u b j ct(s )
e

– Pos s ib le s u b j ct clu s te rs note d at th e s am e tim e
e

– At th e tim e , s taff cou ld not s e e th e e ffe ct of tagging,
as s ys te m was not b u ilt

– R u le s d e ve lop e d as a grou p e xe rcis e with d aily
fe e d b ack and d is cu s s ion as tagging p rogre s s e d

Id e ntifying clu s te rs ...

26

h ttp :/ b e ta.vam .ac.u k/ age / / e m b e rs h ip
/ p m m
h ttp :/ b e ta.vam .ac.u k/ age / kim ono
/ p k/

Th ings to take away...

 Tagging offe rs a p ragm atic m e th od for aggre gation
with ou t cu m b e rs om e s ch e m as

 It is lab ou r inte ns ive

 C las s ification is not as s im p le as it ap p e ars . Tagging
ne e d s b alance b e twe e n accu racy and “good e nou gh ”

 Tagging is ve ry s u b j ctive
e

 Pe op le as s ign im p ortance accord ing to th e ir knowle d ge
of a s u b j ct
e

 D e ve lop ing tagging with ou t s e e ing th e re s u lts cre ate s
anxie ty

e m ail: a.le wis @ vam .ac.u k
And re w Le wis
Linke d In: h ttp :/ b it.ly/ O p AN F
/ h

h ttp :/ b e ta.vam .ac.u k
/

Th ank you
M u s e u m s and th e We b
7 Ap ril 201 1

Mixing it up: Developing and implementing a tagging system for a content-rich website which uses aggregated content from multiple sources.

More Related Content

What's hot (19)

Similar to Mixing it up: Developing and implementing a tagging system for a content-rich website which uses aggregated content from multiple sources. (20)

More from Andrew Lewis (20)

Recently uploaded (20)

Mixing it up: Developing and implementing a tagging system for a content-rich website which uses aggregated content from multiple sources.

Editor's Notes