Welcome to the Q&A Forum
Browse the forum for helpful insights and fresh discussions about all things SEO.
Google is wishy-washy on the answer to this, but Rand's answer is the one I lean on - https://moz.com/community/q/moz-s-official-stance-on-subdomain-vs-subfolder-does-it-need-updating
I think that's a great thread to review.
https://moz.com/community/q/moz-s-official-stance-on-subdomain-vs-subfolder-does-it-need-updating
It's not really eCommerce - you're only facilitating transactions elsewhere, either through a downloaded app or another site. You're more of a portal.
With that said - and I mean this without any tone - your site model is really, really common. I had a client in your market for a short time, and they looked very similar to you as well. In doing competitive research I saw many others with giant collections of pages, regurgitated "descriptions" and no real expertise, differentiation or personality.
The recommendation I made for them, which they didn't take, is prove to Google you're not a quickly pulled together collection of advertisements. Instead, use your words to show why you should be Top 5 in your market. So I would recommend you change focus and start doing something new. The whole affiliate space might have something to be concerned with here.
(not to mention, you seem to have some clunky content on some of the pages I saw, similar to stuff I see from cheap content houses)
"Make her part of the solution, rather than being a problem." I like that a lot.
She needs to understand that focusing solely on one (or a few) keywords is chasing your tail. Google makes hundreds of changes a year - she's going to get knocked down, it's part of the game. I always prefer to win more keywords by doing broader work. Focus on the traffic from all your terms, than the rankings of one.
I've been there tons of times. Try a sit-down with her, talk about the change of focus and expectations. Show her analytics reports to back up a new direction. Create new KPIs. Try to get her excited about a new definition of success. That's what I would do 
Make her part of the solution, rather than being a problem.
It's really a design exercise. You just put your dynamic span tags in the HTML positioned on top of the image.
There's a number of resources - this is one: http://www.the-art-of-web.com/css/textoverimage/
Good luck!
I actually request a number of inquiries (with parameters). I want to see that the attempt not only is made, but the follow ups are there, the outreach is righteous, and it's quality. Most great links are really hard to get, so I'm happy that the best foot was put forward. Easy to track in something like Buzzstream.
That number is estimated based on the talent of the builder and the vertical and/or past experience. It's raised as the team member gets better... like weightlifting.
Headings are used to structure content. Use them logically like you would in a word processor. I always refer to a Wikipedia entry as a great, logical use of headings.
Search engines started using headings as a signal because, well, it makes sense. Surely they knew in the early days that often pages wouldn't be designed with a clear understanding of headers. Then SEOs started exploiting on these headers because they knew they were signals.
That brings us to today... I believe Google is much less focused on these headings unless implemented logically. So my recommendation is to use them logically. Don't think of them from an SEO point of view, but from a logical document flow. The SEO will come from that.
So to answer your question, since I don't know the page, does it logically make sense to make them an H3 tag in terms of your overall document? You might be able to answer your own question if you think I'm correct
Google still internally values / devalues the links they discovered based on their algos. I think in 2011, when penalties and WMT notifications started to pop up, many SEOs forgot that Google still does a good amount of this internally. They'll react when it's egregious.
If you see pure spam in your backlink profile, Google suggests you go ahead and disavow. If you see otherwise good domains that are irrelevant (which it sounds like you do), then there's no reason to panic until Google misreads the intent and thinks you're trying to game the algorithm. If you're not doing that, you're probably in fine shape continuing to build good relevant links and creating the most relevant and valuable website possible.
egregious
It can be done, but I would recommend the HTML overlay the graphic. Then users have the ability to click-and-dial direct in mobile.
Great topic. I wrote a post about this a few years ago - love to hearing the answers for 2016.
2 old school tactics that I think do still work are:
Title tag optimization - I don't mean positioning of the keyword, count, etc., but the power of keywords in the title is still pretty remarkable considering most things that are easily spammable were ultimately deprioritized by Google.
PageRank sculpting - "Crawl Budget sculpting" may be apt, but sculpting the bots path is still incredibly valuable. We do it all the time in our technical work, and there's often a measurable result. In my opinion, the more Google improved their ability to crawl the entire site, the more they learned the reasons to lean away from some sites' sections. Giving only the proper paths to Google keeps it simple for them.