What is the importance of areas in OSPF?

Ospf is a link state routing protocol, so it process a lot more information than any distance vector or hybrid protocol. 

it never gives updates to its neighbors , instead it advertise different lsa so it trigger more cpu utilization . we have periodic lsa flooding and we can also have triggered updates so if our network is quite big more like 50 to 100 routers or more then if we have ospf running , there are high chances of high cpu utilization. to avoid this ospf instead has the concept of different areas . The links state behavior is only bounded to that particular area. Outside that area the border router cannot advertise elaborative link state information so cpu utilization is covered. Another benefit is it’s easy to troubleshoot as each area will have its own network so if there is a problem in an area it’s easy to diagnose than looking for each router in an single area.

Few area types 

area 0 or backbone (without it , two none backbone areas cannot communicate)

stud,tottaly stub,nssa and totaly nssa (these area types are for specail cases to reduce cpu ulization or lsa flodding to another level)