In what scenario would the cities on the coast be the only ones that matter?
Look at New Hampshire. It's worth a measly 4 electoral votes, only 0.74% of the electoral college, and still a fairly small amount of the total swing votes reasonably considered to be in play. And yet we saw quite a lot of campaigning and canvassing going on there because it was a very competitive state in a very competitive race (which most presidential races these day are) and at that point every bit that could be secured mattered.
In this election the popular vote was very close, off by only around 1%. Even the counties that are the deepest shades of red or blue contain some often significant fraction of people who could be convinced to vote the other direction. This is especially true in the modern era where people are easily exposed to information well beyond their home town. If one candidate managed to successfully gain votes among this minority over the other they could have very realistically turned the election. This is why a lot of real minority blocs are important to campaigns today; the majority is not considered "enough." Even if they currently only matter if they're in swing states and can tip the election.
The largest cities would have the biggest low hanging fruit, but the campaigns will reach diminishing returns where they feel like they can cause more of a shift in preference and turnout in increasingly less populated areas. Especially if they can reach large swaths of them at a time by appealing to their more unifying values.
This is very different from the winner-takes-all method employed in most states, where the people who voted for the non-majority party literally contribute nothing to the outcome, and in the large number of states that are very entrenched and slowly change affiliation they can know ahead of time that their vote is a waste. Appealing to the minority voters in those states is a waste of time.