There's never one perfect solution, however here's the bigger issue.  Some people hear "flat" and they take it to the extreme. Which is a terrible concept in 2011.

If you go too flat, you muddy up the proper group relationships.  This is where Siloing comes in.

In my presentation at SMX Advanced this week, one of the many methods I recommend for "sustainable SEO" is to group your content, and reinforce that group relationship in URL structure, then with breadcrumbs, and finally with section-level navigation, where all the pages in that section have a link to all the other pages in that section, but where that specific sub-navigation is replaced or disappears as appropriate when you leave that section.

If you've got more than a handful of pages in a section, you should definitely go deeper.

The trick is knowing how wide, how deep to go.  It's an art as much as a process studying site data over time.

Another factor is the competitive landscape for a particular niche market.  The more competitive, the more important this concept becomes.

I'll just add that Yahoo closed the Yahoo Directory down last year.

Actually the correct deeper answer based on both Google policies and SEO best practices is as follows:

1. It is directly against Google's terms of service to attempt to rank multiple sites for the same phrases.
2. When you have more than one site that contains content that directly competes against any other site, whether its a site you own or someone else owns, or even other content on your own site, Google's multi-algorithm system attempts to determine which site deserves the higher ranking for a particular phrase or search query.  In that process their system attempts to then determine whether any of those shouldn't even be indexed, let alone show up in search results.
3. Based on these considerations, any of your content could quite possibly suffer from either a loss of position it should otherwise deserve, or even have some or all of its content deindexed.  And in a worst case scenario, you could be penalized as well.

SO - the only issue then is this - WHY would you want multiple sites? Do any of the following reasons match your vision? If so, then you CAN have multiple sites IF they are done properly.

A) If you've got a big active brand, with a lot of customers/clients, it can help to create multiple sites often including:

• Corporate Site
• eCommerce Site
• Careers Site
• Community Site
• Charitable Giving Site
• Customer Support Site

B) If you have specific separate and quite distinctly different service or product offerings, you can create multiple sites so that the very different topical intent of each site is kept uniquely refined in that specific funnel and doesn't "pollute" or "dilute" the umbrella topical focus of each niche.

C) If you have an eCommerce site (where intent is online sales) you may have a desire to have a separate community or blog site (where intent is informational) as another way to keep the "intent" funnels cleanly separated.

NOTE:

It is VITAL that you understand the concept that when executed properly, multiple sites are very useful. However, these need to factor in the following:

1. Every site needs to be able to pass the "5 Super-Signals" test:

• Quality
• Uniqueness
• Authority
• Relevance
• Trust

In regard to the above, content needs to be truly unique across each site. While you can have similar content specific to your brand identity, and even some similarity about the umbrella topic of your product or service offerings, this needs to be done in a way that does not violate the "multiple sites for ranking domination" except as it relates to your brand (as opposed to generic non-brand product or service offerings).

2. Each site needs to have a LEGITIMATE business case reason for its existence not considering SEO - the "why this site exists" question needs to pass muster.

3. Every additional site you create requires its own consistent quality effort, as well as trustworthy off-site reinforcement. If a proper concerted effort cannot be maintained over the long-haul on multiple sites, it is much wiser to go with one single unified site.

Always place schema markup directly in the position on the page where you want the content to appear if it's content specific - wrapping it around that content. So if your business name and address are in the main content area, that's where you place the schema code.  It's literally a wrapper just like a CSS div would be, or an old-school HTML table, but not for display purposes on-site.

EDITED 11/14/2013 based on a question from Oliver (below) regarding situations where markup is located in the "head" area of the page:

Exceptions to "in-body" markup:

As is the case with any structured markup solution, there will, from time to time, be cases where certain, specific elements go in the "head" section of the code.  Anything that applies to an individual page in its entirety, and does not limit itself to an element of content within the page does, in fact, belong in the "head" area of the page code.

There's some disagreement in the industry as to whether the pipe symbol or hyphens are best - either way, one of these two would be recommended for readability purposes.  This is especially valid when you've got more than one keyword phrase.

Product Name | Alternate Product Name | Company Name

Just offering my opinion.  There is no such thing as "concrete proof" that can't be disproven in this case due to the complexity of SEO.

Every factor is just one among many.  So a site that has "proper" URL syntax can easily and readily outrank and outperform a site that doesn't if enough individual factors across the whole spectrum are strong enough.

Conversely, A site that has numeric URL structure and "non-ideal" syntax can also easily and outrank / outperform a site that has "proper" URL syntax if that site has enough strength from other factors to outweigh the "proper" structured URL site.

Anyone who has a case study claiming otherwise is not acknowledging how complex the reality of what we do is, and how any sub-group of signals can be so strong as to far outweigh any other sub-group of signals.

JavaScript is one of several technologies that offers severe limitations in search engines and their ability to properly see content, then just as important but often overlooked, properly and cleanly evaluate that content from an SEO perspective.

Specific considerations:

• Google does a "fair" job at discovering content passed through JavaScript (either on-page or at the code level)
• A "fair" job means it's hit and miss as to whether their system can actually find that content
• Whatever content the Google system CAN find via JavaScript is NOT necessarily able to be used to properly evaluate content intent, focus or relationship to other content

So - the best practices recommendation is if you want/need content to be found and properly evaluated by Google (or Bing) do NOT pass it through JavaScript.

And also, if you want to HIDE content from Google, don't assume you can successfully do so via JavaScript either.

As for PHP, its the most widely adopted and utilized web programming language out there.  The language by itself is essentially SEO neutral.  It's all in how a programmer utilized PHP that matters.  In the hands of a programmer that either truly understands SEO or collaborates closely with an SEO expert (who also understands the limitations/pitfalls that can arise with "bad" (SEO-unfriendly) PHP coding, it's a great language.

Eric,

Unfortunately you're in a very difficult position.  Personally, I would never proceed with a client who can't even define their own audience. They're quite likely going to change their mind often, and if you get involved this early on, will just as likely be overly demanding and play the "needy victim" role.  It's a mess.

Having said that, if you insist that you absolutely must work with them, the best approach might be to choose four, five or six topics you think are appropriate, run them through the Google Keyword tool, export each result set and pass them an Excel spreadsheet with each result set being in a separate tab.  Then, help them understand what the columns mean (competitive, search volume, etc) and let them chew on that data.  Explain that you need them to decide which phrases to go with, but that you can help them refine it down a bit.

Be very careful though to not get sucked into an endless hours vortex!

I'd caution against assuming it's meaningless, but only from the perspective of not HTML specifically.  Instead, if a page is bloated with code at the source / crawler bot level, that can definitely have a negative impact, because it can cause topical confusion.  Whether the code is essentially gibberish to the human eye, or javascript with a lot of text that describes functions, calls and parameters.

I see such problems quite often on large sites that try to do a lot at the code level.

Well it depends. Is there only one BMW or are there several? If there is only one, then yes - cross link all the luxury detail pages.  If there are several, then that's the level for cross linking detail pages, even though it's so deep.  If that's the case though, you'd better get inbound links pointing to the parent luxury category page.

And in any regard, don't just have a bunch of links on those category pages - have descriptive paragraph content focused on that category's primary topical focus.

It's ultimately irrelevant from an SEO perspective - so it comes down to click tracking.  If you have different actions, you'd want a unique shortened URL for each action.

I'd also suggest considering setting up your own URL shortener, rather than relying on 3rd party.  As great as Bit.ly is, and as much as Twitter offers their own, the only way to ensure a short URL that's generated is always going to be around is to have your own.

If you mean - if we get an inbound link pointing to the subdomain, will the root domain benefit?" the answer is yes to a certain extent.  How much will depend on how closely related the subdomain topical focus is to the root domain, as well as how strong the link is from the perspective of the source site.

If you mean - if we point a link from within a subdomain to a page on the root domain, will the root domain benefit?" Then the answer is also yes, but how much depends on many other factors. And those links will carry less value than most quality inbound links coming from other root sites not in your control.