- how do you do it for SEO when you don't know what Google is lookin' at?
I don't give a rat's rear what Google is looking at. Except for the pages I don't want them to index.
I care about what (converting) human visitors are looking at. And guess what? Google does it's thing whatever that might be and seems to like what they see anyway.
SEO is one of:
* not putting up/removing barriers to SE bot crawl and current rendering ability.
* a recovery program for folks who screwed up SEO-ing their site.
- how do you do it to make a pleasing website experience?
Make it whatever will get your target audience(s) to:
* stay a while;
- how do you do it maximize visitor engagement?
* rinse, repeat;
- how do you do it to maximize ad revenue?
Sell direct and don't sell what/how everyone else is selling.
Outline of doing it my way: Direct ad sales, differently
- how do you do it to maximize sales?
eCom I don't do. I have opinions but little experience.
- how do you do it for speed yet maintain features and functions?
Strip, compress, cache, eliminate/minimise third party calls, progressive enhancement/lazy load, etc.
Less sharing of server, possibly going dedicated.
Consider HTTP/2. Note: it optimises very differently from HTTP1.1.
If primarily US consider simply hosting geo-centrally.
Investigate plus/minus of CDN (not necessarily a good solution).
Really, does anybody understand it. Lots of people talk a good talk but we know how people bullshit. They are fguessing at it.
I fguess - although I mostly call it an educated hypothesise - and then I test.
I'm rather like Google in that regard: always testing. A web business is actually quite a complex intermeshed set of components; we like to separate into components for ease of discussion but changing one, i.e. SEO, without considering another, i.e. UX, is soon problematic. Too many specialists running around without strategic oversight make a mess.
Note: so-called SEOes running around doing stuff other than actual SEO in the name of SEO make a catastrophic mess:
SEOes rush in where fools fear to tread
Like most small businesses webdevs wear too many hats, have too much on their plates, have too little time. I simply keep working towards making 'things' qualitatively better and leave the quantitative commodification race to the bottom to everyone else. Plus play with 'stuff', because: why not? it's fun!
Best mobile advice I have to share:
Always stress test your site on older phone such as iPhone 4 and an Android equivalent, i.e. Galaxy Nexus, HTC One, Droid Razr Maxx, over a marginal/intermittent 3G connection. If you can offer a decent usable site experience to a visitor on that basis... it is easy to forget that not everyone has either the latest device or the best connection. Even in big well connected cities there are tunnels, parking garages, subways, elevators et al.