alot of this is interconnected, so apologies for repeating myself; they do help build a few themes that stand out.
bad news: somethings are really holding you back.
good news: most, including biggest, can be fixed with one or a few short steps.
good bad news: it all depends.
google's mobile-friendly test returns page partially loaded results
this is the most pressing item i see: googlebot smartphone being blocked by cloudfront's robots.txt.
this affects a few things, most importantly search. its not getting any of your image assets on your index. and i'm assuming from any/all documents that you house there as well.
its also throwing redirection errors on two images, one is ga. the analytics image error could possibly be affecting mobile search/other things too...
everything else below is not emergency except for the alt="" attributes on <img />'s. :)
use <article> or <section> for the wrapper element of "Recently Completed Features", and all of the features with an image and read more links; whichever one is for syndicated content.
use microformats, xfn, schema.org, microdata, and json-ld anywhere possible....dates, events, companies, places, people, etc.
anchor element naming conventions: the text around and in <a> elements should describe where/what the link is pointing to/references.
use progress for show completion bars.
use figure, figcaption for the features with an image and read more links.
<meta> element game: use dublin core, description, robots, keywords to name a few create good titles and snippets in Search Results
i know keywords and description is not seo juice, no snake oil here. but they are still used by google and other search engines for various purposes. like what is shown in serps.
set up open search and corresponding open search documents to work with <link type="search" />
add a lang="en-US" <html>; its better for accessibility, and also enables various functionality in crawlers, bots, search engines.not sure why markdown looks like this...
<img />
no alt attributes
naming conventions
example:
<a href="/news/archives/2017/may/15/police-riot-gear-budget/" title="Police departments across the country are spending millions on riot gear" class="article__image">
<img src="https://d3gn0r3afghep.cloudfront.net/news_images/2017/05/15/RiotCov.jpg.900x300_q85.jpg"></a>
could be:
<a href="/news/archives/2017/may/15/police-riot-gear-budget/" class="article__image">
<img src="https://d3gn0r3afghep.cloudfront.net/news_images/2017/05/15/police-department-in-full-riot-gear.jpg" alt="Police departments across the country are spending millions on riot gear" ></a>
swapping the title attribute on the parent anchor elements should be easy enough, as shown above; i like the tooltips but they're not good for accessibility.
why you should add schema.org for mobile image search
use schema.org's MediaObject with associated caption linking (internally) to NewsArticle containing said MediaObject
you could wrap most of what i've said above together into a few patterns.
run them through google structured data testing tool as well other validators if you're going to use more than schema.org.
swap time element for the datetime element
svg logo duplicates title attribute content in another attribute in some places, which returns as "MuckRock Logotype A boldfaced"; change that to whatever the link is to, pretty sure that's index...
Without access to webmaster tools aka search console, can't do in-depth analysis
also no access to Crawl Errors Report
desktop overall web performance is optimal. i don't see these being real issues, although i didn't check/compare image sizes, because they didn't come back as issues/lazy.
add an expires header
compress assets/min images
reduce dns lookups
ga is largest load time; easily swap that for the load after pageload technique.
excellent structure and naming conventions.
because i'm obtuse, i'd change named months 'feb' to numbers '02'
submit url to google search or via google search console
use canonical URLs
documents/images are part of this/also using rel="canonical" when possible with documents/images.
hosting?
you want to own your content, including all your assets. 3rd party hosting gets your image and document google juice. they also can have certain limitations, particularly with naming conventions and optimizations. also, in your case they're blocking bots entirely. if you are using one for your documents, prety sure its document cloud, you're also losing all of that content google juice and documentcloud/cloudfront is getting all of that google juice.
providing transcripts/captioning for media provides content for seo and content for accessibility. win win.
media includes pdfs...
that said, i'm sure you handle massive amounts of documents so this is all so much easier typed than done. i stand by it entirely, but realize how difficult/impossible it can be with existing systems.
rel="canonical" on <link />/images/documents
.htaccess can be used here
use alt attributes on every <img /> element that you use. there are rare cases when not to, but they won't come up. seriously, every single user that comes to your site with a screenreader gets nothing from alt-less images, using it makes your site accessibe. bonus: they're used for content/indexing for serps and image search.
figure/figcaption use were appropriate.
the words around a link are important but the most important words are inside the link.
"Read More" and "Learn more" links are hurting your internal linking game, and hurting your indexable content game.
make replacements for both that form solid connections between index and page references, particularly around prominent pages/sections. possible solutions:
for learn more, just make the intro paragraph/sentence the link. also boosts usability = bigger click space
for read more...make the last sentence the link?
create sitemap/submit to google/search engines
create clustered keywords plan based off queries using keywords results in similiar urls in serps
google keyword planner for long-tail keyword variations
read this google search engine optimization starter guide (PDF)
seriously read this google search engie optimization starter guide (PDF)
seo is not fun to read about, but they literally tell you step-by-step instructions to follow, from which you can just build
your content strategy/ai/etc. around.
dig into your logs and see what/where your users are coming from; not flawless, but including the majorites/outliers of
your users into your content strategy tends to play extremely well with each other.
do all of this in bing too. yandex and china's search engine, etc.