Drop stock photography for a typography-led design, rewrite the closing CTA

The guide previously leaned on generic stock photos of strangers for each
step, which didn't match reground.org's actual visual identity (color,
type, and voice carry the brand everywhere else on the site — no stock
imagery appears anywhere but the author's own photo). Removed all stock
photos; the only photo left is the real author headshot on the cover and
closing page, which also got a real cleanup (tighter crop, better contrast,
reduced glasses glare) and a bug fix (it was rendering as an oval instead
of a circle since width/height were never equal).

Since nothing forces per-step pages anymore, the PDF's h2 sections now flow
continuously instead of each getting its own page-break — a short guide
with no photos to fill a page reads better as continuous pamphlet-style
prose than as a sparse slide-per-page export. Each step keeps a short
pull-quote (verbatim line from its own paragraph, styled in the brand
blue/teal) for visual rhythm instead. Net effect: 6 pages -> 3 pages, no
blank/orphaned space. The pptx keeps one section per slide, since that
format is for narrated screen-share recordings where the presenter's voice
fills the space a printed page can't.

Also rewrote the closing CTA: "Get a Free Strategy Session" was generic
coaching-industry boilerplate that could belong to anyone's funnel. The new
copy and button use the guide's own vocabulary instead.

README.md

@@ -36,8 +36,12 @@ make partners/<slug>.pptx        # a specific partnership's deck
 make slides                      # template.pptx + every partners/*.pptx
 ```
 
-Pandoc turns each `##` heading into its own slide, matching the "one step per
-page" layout of the PDF. If you want scripted talking points per slide, add a
+Pandoc turns each `##` heading into its own slide. The PDF no longer mirrors
+this one-section-per-page layout — since the guide has no per-step photography
+to fill a page, its sections flow together into continuous pages instead
+(see Notes below). The slide deck keeps one section per slide on purpose: it's
+built for a narrated screen-share, where your voice fills the space a printed
+page can't. If you want scripted talking points per slide, add a
 `::: notes` ... `:::` fenced div under a heading — pandoc exports those as
 speaker notes rather than slide body text.
 
@@ -57,7 +61,7 @@ cp template.md partners/<slug>.md
 ```
 
 Edit the new file's title (front matter) and rewrite the opening/closing copy to bridge
-your angle with the partner's audience — the 3 Step sections and their photos are
+your angle with the partner's audience — the 3 Step sections and their pull-quotes are
 generally reusable as-is. Then:
 
 ```sh
@@ -66,9 +70,16 @@ make partners/<slug>.pdf
 
 ## Notes
 
-- The photos in `images/` (4 extracted from the original PDF, plus `intro-reader.jpg`
-  sourced from Pexels) are intentionally generic/illustrative (not specific to any one
-  partnership) and don't need to change per customization.
+- `images/author-photo.jpg` is the only photo in the guide, used on the cover and
+  closing page. The guide deliberately doesn't use stock photography anywhere else —
+  matching reground.org, which carries its identity through color, type, and voice
+  rather than illustrative imagery. Each Step section instead gets a `.pull-quote`
+  div (see `style.css`) pulling one line verbatim from that section's own paragraph,
+  which gives the page visual rhythm without adding a picture.
+- The PDF's `h2` sections flow continuously (no forced page break) since there's no
+  photo to justify each Step having its own page. Only the cover and the closing CTA
+  force their own page break — a short guide with unbroken flow reads like a real
+  pamphlet rather than a slide deck exported to PDF.
 - `template.html` renders the cover (title + byline + author photo) from the document's
   YAML front matter — you don't need to hand-write that part in the body of each `.md`
   file, just update the front matter.