I must say that I do not agree and the problem is that you choose to use an ester as starting material. The ketone-product, that is formed from addition of the grignard-reagent, is much more reactive (since it is more electrophilic) than the ester and once it is formed it will preferentiably react with the grignard. This will give a tertiary alcohol as product instead of the ketone (you will get a mixture of ketone and alcohol if you use 1 molar equivalent of grignard).
The solution to the problem is to use something which is more reactive than the ester, e.g. an acid chloride or to use an aldehyde and then afterwards oxidize the formed alcohol to the ketone.