I believe you need something like
select count(*)
from customer c
where c.key = 'lastName'
and c.value in ('Johnson', 'Brett', 'Jack')
so you can count all the customers with last names including Johnson, Brett and Jack.
With you current data model, if you want to find all people with a specific firstName and lastName you must join two instances of customer table as
select count(*)
from customer c1
join customer c2 on c1.user_id = c2.user_id
where c1.key = 'firstName'
and c2.key = 'lastName'
and c1.value = 'John'
and c2.value = 'Doe'
and it goes more complex if you need to search upon more fields.
By the way, why do you use such a data-model in an RDBMS? If you have an schema-less model, why don't you use a NoSQL database such as Mongo?
