(1) Make valuable recommendations (2) Offer me a better price than any website (3) Save me hours upon hours of wasted time
When it's more than the website, I do a cost-benefit analysis and decide whether the value I'll derive from paying a slightly higher price for amenities is worth it or not.
And, I get the benefit of the travel agent's expertise, as you say, who knows a little bit about me and what I care about. Sure I can do the research myself (and I will often do so), but it's nice to get all this stuff for basically nothing.
(1) I'd rather the aggregate recommendations of a million people than the personal opinion of just one in most of these situations - you want the statistical mean experience, because a travel agent can be given special treatment or just have had a lucky streak placing clients in an otherwise seedy place.
(3) Hopefully another thing tech can fix. I mean, yeah, you have the rare extremely special case circumstance you can't easily break down into a power search, so they have a niche, but as a general tool for people looking for places to stay, you really don't need another fleshy body directly involved with the seeker.
(1) How many online reviews are fake? How many are written out of spite? There is a level of trust you have to give someone: either an online reviewer or a person.
(3) Having a human to work with is usually better service than not. But it really depends on the situation.
Edit: I'd like to mention that I am not an agent of any sort. I do more business online than anyone I know, but I also realize the pros/cons of both sides of the spectrum.
That is why I quantify pretty explicitly a million - I'd never consider any sampling below, say, a hundred reviews, and even then you have to accept the inherent inertial bias that anyone content with their experience won't review it, and disproportionately people will complain about a negative one than praise a great stay. You need enough aggregate results to smother false reviews or non-systemic bias.
> (3) Having a human to work with is usually better service than not. But it really depends on the situation.
It is really situational, but one critical aspect of the information age is a litmus test on every human service position on whether or not actually having gassy meat on one end of the transaction is really worth the comparatively huge costs of keeping said meat bag alive and happy. And it is a personal test - which is why most industries that have been displaced by the Internet still exist to serve the population that values that interaction.
