I've seen and read several articles on how store designers and owners can create pseudo faceted navigation (also sometimes called guided navigation) using tags to mimic sub-categories. This may work for basic stores. But as someone who has managed multi-million dollar ecommerce websites in Canada, I can tell you flat out that's a hackish approach that ignore the real use scenario.
What Shopify desperately needs is a faceted navigation system that's driven by discreet product attributions (metadata). For example, a website selling car radios should have very detailed and rich product information. The "Tech Specs" list for a product would be very long and detailed. Remember: A lot of e-commerce activity online is about researching products, so this type of information is valuable to the pre-purchase phase. This data (colour, size, power capability, EQ bands, etc., etc) should be stored in metafields as discreet product info. Tags are not appropriate for this type of data storage because they are not key-value paired. In the case of our car radio example you might have "Preamp Outputs: 4 channels", so you have a key of "Preamp Outputs" and a value of "4 channels". Tags don't come close to doing this.
With that data properly stored in the metafields, the faceted navigation system should display menus on the left side of the page based on the current collection displayed. These menus are driven by the metafields, not tags. So, returning to our example of a store that sells car radios, if you're currently viewing a collection for car stereos that have GPS navigation builtin, I should be able to see a navigation menu on the left side that allows me to filter the product listing based on all known discreet attributions for products in that collection. For example, product metadata such as colour, iPhone compatible, power outlet > 100 watts, etc., etc.
You can't power this kind of detailed product navigation with simple tags.