diff --git a/app.py b/app.py
index 01f2965..c98804c 100644
--- a/app.py
+++ b/app.py
@@ -12,11 +12,18 @@ import pandas as pd
 import report
 import growth
 import visitor_history
+import monthly_report_notes
+import follower_directory_data
 
 # Overridable via env var so the same code runs unmodified in Docker/internal deployments —
 # defaults to this machine's path for local dev.
 DATA_DIR = os.environ.get("LINKEDIN_DATA_DIR", "/Users/mckinleyblack/Documents")
 UPLOAD_DIR = os.path.join(DATA_DIR, "linkedin_analytics")
+ASSETS_DIR = os.path.join(UPLOAD_DIR, "assets")
+
+# Loaded once at startup, not re-parsed per callback — re-run the app to pick up new
+# assets/follower_directory.html entries, same as every other data source here.
+DIRECTORY_DATA = follower_directory_data.load_directory(ASSETS_DIR)
 
 # LinkedIn's export only gives daily gains, never an absolute follower total. This is
 # the one manual anchor point: the count going into the very first day this export
@@ -1347,6 +1354,7 @@ def refresh_data():
     TAB_CONTENT_MAP = {
         "overview": overview, "content": content_tab, "followers": followers_tab,
         "visitors": visitors_tab, "competitors": competitors_tab, "growth": growth_tab,
+        "reports": reports_tab,
     }
 
 
@@ -2023,6 +2031,46 @@ def build_growth_tab():
     ], fluid=True, className="pb-3")
 
 
+def _reports_month_options():
+    """Only months with curated narrative notes are selectable here, not every month
+    with export data — this tab is about the campaign's written progress reports, not
+    a generic timeframe filter (that's what the sidebar's PDF control already does)."""
+    import calendar as calmod
+    opts = []
+    for key in sorted(monthly_report_notes.NOTES.keys()):
+        year, month = (int(p) for p in key.split("-"))
+        opts.append({"label": f"{calmod.month_name[month]} {year}", "value": key})
+    return opts
+
+
+_REPORTS_MONTH_OPTIONS = _reports_month_options()
+_REPORTS_DEFAULT_MONTH = _REPORTS_MONTH_OPTIONS[-1]["value"] if _REPORTS_MONTH_OPTIONS else None
+
+reports_tab = dbc.Container([
+    html.Div([
+        html.Div([
+            html.Label("Reporting Month", style={"fontSize": "0.75rem", "fontWeight": "600",
+                                                   "marginBottom": "4px", "display": "block"}),
+            dcc.Dropdown(
+                id="reports-month-select",
+                options=_REPORTS_MONTH_OPTIONS, value=_REPORTS_DEFAULT_MONTH,
+                clearable=False, searchable=False, style={"width": "260px"},
+            ),
+        ]),
+        html.Div([
+            html.Button("Download PDF", id="reports-download-btn", n_clicks=0,
+                        style={"padding": "8px 16px",
+                               "background": "linear-gradient(90deg, #1b3a5c 0%, #2d6a9f 100%)",
+                               "color": "white", "border": "none", "borderRadius": "8px",
+                               "fontSize": "0.8rem", "fontWeight": "700", "cursor": "pointer"}),
+            dcc.Download(id="reports-download"),
+        ]),
+    ], style={"display": "flex", "justifyContent": "space-between", "alignItems": "flex-end",
+              "marginBottom": "1.25rem"}),
+    html.Div(id="reports-content-area"),
+], fluid=True, className="pb-3")
+
+
 refresh_data()
 
 
@@ -2040,6 +2088,7 @@ NAV_ITEMS = [
     ("visitors",     "Visitors",     visitors_tab),
     ("competitors",  "Competitors",  competitors_tab),
     ("growth",       "Growth",       growth_tab),
+    ("reports",      "Reports",      reports_tab),
 ]
 TAB_CONTENT_MAP = {key: content for key, _, content in NAV_ITEMS}
 
@@ -2050,6 +2099,7 @@ TAB_HERO = {
     "visitors":    ("Visitors",    "Page views and unique visitors to the TXproducts LinkedIn Page, broken down by device and section."),
     "competitors": ("Competitors", "Benchmarking TXproducts against tracked competitor pages on LinkedIn."),
     "growth":      ("Growth",      "Historical follower growth, tracked weekly against your own rolling best."),
+    "reports":     ("Reports",     "Monthly written progress reports for the 3-month LinkedIn campaign, combining dashboard data with the Follower Directory."),
 }
 
 GROWTH_METRIC_EXPLAINER = html.Div([
@@ -2505,6 +2555,18 @@ def _sync_content_charts(selected_month):
     )
 
 
+def _build_report_ctx():
+    """Shared by the sidebar's generic PDF control and the Reports tab, so both build
+    the exact same ctx_data dict for report.build_report_pdf/compute_period_stats."""
+    return {
+        "D": D, "posts": posts, "comp_ordered": comp_ordered,
+        "current_total_followers": current_total_followers,
+        "daily_followers_figure": daily_followers_figure,
+        "std_fn": _std, "TMPL": TMPL, "FORMAT_COLORS": FORMAT_COLORS,
+        "growth_snapshot": GROWTH_SNAPSHOT, "directory_data": DIRECTORY_DATA,
+    }
+
+
 @app.callback(
     Output("report-download", "data"),
     Input("generate-report-btn", "n_clicks"),
@@ -2512,19 +2574,168 @@ def _sync_content_charts(selected_month):
     prevent_initial_call=True,
 )
 def _generate_report(_n_clicks, timeframe):
-    ctx_data = {
-        "D": D, "posts": posts, "comp_ordered": comp_ordered,
-        "current_total_followers": current_total_followers,
-        "daily_followers_figure": daily_followers_figure,
-        "std_fn": _std, "TMPL": TMPL, "FORMAT_COLORS": FORMAT_COLORS,
-        "growth_snapshot": GROWTH_SNAPSHOT,
-    }
-    pdf_bytes = report.build_report_pdf(ctx_data, timeframe)
+    pdf_bytes = report.build_report_pdf(_build_report_ctx(), timeframe)
     label = next((o["label"] for o in report.report_month_options(D["fts"]) if o["value"] == timeframe), timeframe)
     safe_label = label.replace(" ", "_")
     filename = f"TXproducts_LinkedIn_Report_{safe_label}.pdf"
     return dcc.send_bytes(lambda buf: buf.write(pdf_bytes), filename=filename)
 
 
+def _narrative_section(title, items):
+    if not items:
+        return html.Div()
+    return html.Div([
+        html.H5(title, style={"marginTop": "18px", "marginBottom": "8px", "color": "#1b3a5c"}),
+        html.Ul([html.Li(item, style={"marginBottom": "4px"}) for item in items],
+                style={"fontSize": "0.85rem", "paddingLeft": "1.1rem"}),
+    ])
+
+
+def _composition_table(counter, header_left, top_n=None):
+    """No truncation by default, sector/seniority/function are a small, fixed taxonomy
+    (see project memory's classification categories), and country is uncapped by
+    composition_stats() for the same reason, so showing everything keeps the table's
+    own sum matching the cohort total instead of silently dropping tail categories."""
+    items = counter.most_common(top_n) if hasattr(counter, "most_common") else list(counter)[:top_n]
+    return dbc.Table([
+        html.Thead(html.Tr([html.Th(header_left), html.Th("Count")])),
+        html.Tbody([html.Tr([html.Td(k), html.Td(str(v))]) for k, v in items]),
+    ], bordered=True, size="sm", style={"fontSize": "0.8rem"})
+
+
+def _reports_posts_table(posts_f):
+    """Every post in the period, listed chronologically, not just a top-N by engagement.
+    A monthly campaign report should show the full picture of what was published, since
+    the point is a complete record, not a highlights reel."""
+    if not len(posts_f):
+        return html.Div("No posts published in this period.", className="text-muted")
+    all_posts = posts_f.sort_values("Created date")
+    return dbc.Table([
+        html.Thead(html.Tr([html.Th(c) for c in
+                            ["Post", "Date", "Type", "Impressions", "Eng %", "CTR %"]])),
+        html.Tbody([
+            html.Tr([
+                html.Td(str(r["Short title"])[:60]),
+                html.Td(r["Created date"].strftime("%d/%m/%Y")),
+                html.Td(r["Content Type"]),
+                html.Td(f"{int(r['Impressions']):,}"),
+                html.Td(f"{r['Eng %']:.1f}%"),
+                html.Td(f"{r['CTR %']:.1f}%"),
+            ]) for _, r in all_posts.iterrows()
+        ]),
+    ], bordered=True, size="sm", style={"fontSize": "0.8rem"})
+
+
+@app.callback(
+    Output("reports-content-area", "children"),
+    Input("reports-month-select", "value"),
+)
+def _render_reports_tab(month_value):
+    if not month_value:
+        return dbc.Alert("No campaign months have been written up yet.", color="info")
+
+    stats = report.compute_period_stats(_build_report_ctx(), month_value)
+    notes = monthly_report_notes.NOTES.get(month_value)
+    fts_f, posts_f = stats["fts_f"], stats["posts_f"]
+
+    kpi_row = html.Div([
+        html.Div(kpi("Total Followers", f"{current_total_followers:,}", "as of latest export"),
+                  style={"flex": "1", "minWidth": "0"}),
+        html.Div(kpi("New Followers", f"{stats['new_followers']:,}", stats["period_label"]),
+                  style={"flex": "1", "minWidth": "0"}),
+        html.Div(kpi("Posts Published", f"{stats['posts_published']:,}"), style={"flex": "1", "minWidth": "0"}),
+        html.Div(kpi("Total Impressions", f"{stats['total_impressions']:,}"), style={"flex": "1", "minWidth": "0"}),
+        html.Div(kpi("Total Page Views", f"{stats['total_page_views']:,}"), style={"flex": "1", "minWidth": "0"}),
+        html.Div(kpi("Best Post Eng.", f"{stats['best_eng']:.1f}%"), style={"flex": "1", "minWidth": "0"}),
+        html.Div(kpi("Avg. Eng. Rate", f"{stats['avg_eng']:.2f}%" if stats["avg_eng"] else "—"),
+                  style={"flex": "1", "minWidth": "0"}),
+        html.Div(kpi(
+            "Monthly Growth Rate",
+            f"{stats['monthly_growth_rate']:+.1f}%" if stats["monthly_growth_rate"] is not None else "—",
+            (f"+{stats['new_followers']:,} vs {stats['start_of_month_followers']:,} at start "
+             f"of {stats['period_label']}") if stats["start_of_month_followers"] is not None else "",
+            color=GREEN if (stats["monthly_growth_rate"] or 0) >= 0 else "#d43f3a",
+        ), style={"flex": "1", "minWidth": "0"}),
+    ], style={"display": "flex", "gap": "0.75rem", "marginBottom": "1.25rem", "flexWrap": "wrap"})
+
+    chart_card = dbc.Card([
+        dbc.CardHeader("Follower Growth"),
+        dbc.CardBody(G(daily_followers_figure(fts_f, posts_f), style={"height": "280px"}) if not fts_f.empty
+                     else html.Div("No follower data for this period.", className="text-muted")),
+    ], className="mb-3 shadow-sm")
+
+    posts_card = dbc.Card([
+        dbc.CardHeader(f"All Posts, {stats['period_label']}"),
+        dbc.CardBody(_reports_posts_table(posts_f)),
+    ], className="mb-3 shadow-sm")
+
+    cohort = follower_directory_data.composition_stats(DIRECTORY_DATA, month_value)
+    all_time = follower_directory_data.composition_stats(DIRECTORY_DATA)
+    if cohort["total"]:
+        composition_body = [
+            html.P([
+                f"LinkedIn recorded {stats['new_followers']} new followers in {stats['period_label']}. "
+                f"Of those, ",
+                html.Strong(f"{cohort['total']} have been identified and profiled"),
+                f" in the Follower Directory so far (a separate, hand-maintained audit of who our "
+                f"followers are, not a second follower count), out of {all_time['total']} audited "
+                f"across the whole page to date.",
+            ], style={"fontSize": "0.85rem"}),
+            html.Div([
+                html.Div(_composition_table(cohort["sector"], "Sector (this month)"),
+                          style={"flex": "1", "minWidth": "220px"}),
+                html.Div(_composition_table(cohort["seniority"], "Seniority (this month)"),
+                          style={"flex": "1", "minWidth": "220px"}),
+                html.Div(_composition_table(cohort["country"], "Country (this month)"),
+                          style={"flex": "1", "minWidth": "220px"}),
+            ], style={"display": "flex", "gap": "1rem", "flexWrap": "wrap"}),
+        ]
+    else:
+        composition_body = [html.P(
+            f"LinkedIn recorded {stats['new_followers']} new followers in {stats['period_label']}, "
+            f"but none have been identified and profiled in the Follower Directory yet.",
+            style={"fontSize": "0.85rem"})]
+    composition_card = dbc.Card([
+        dbc.CardHeader("Audience Composition (Follower Directory Coverage)"),
+        dbc.CardBody(composition_body + [
+            html.P("The Follower Directory is a hand-maintained audit that lags real LinkedIn follows "
+                   "and gets corrected after the fact as more profiles are screenshotted and "
+                   "cross-referenced, so this is a coverage figure that can still grow in later "
+                   "updates, not a final count.", className="text-muted",
+                   style={"fontSize": "0.75rem", "fontStyle": "italic", "marginTop": "10px"}),
+        ]),
+    ], className="mb-3 shadow-sm")
+
+    if notes:
+        narrative_card = dbc.Card([
+            dbc.CardHeader("Monthly Notes"),
+            dbc.CardBody([
+                _narrative_section("Features / Fixes Shipped", notes.get("features_shipped", [])),
+                _narrative_section("Data Quality Issues Found and Fixed", notes.get("data_quality", [])),
+                _narrative_section("Follower Study / Directory Updates", notes.get("follower_study", [])),
+                _narrative_section("Outstanding", notes.get("outstanding", [])),
+            ]),
+        ], className="mb-3 shadow-sm")
+    else:
+        narrative_card = dbc.Alert("Notes for this month have not been written yet.", color="warning")
+
+    return html.Div([kpi_row, chart_card, posts_card, composition_card, narrative_card])
+
+
+@app.callback(
+    Output("reports-download", "data"),
+    Input("reports-download-btn", "n_clicks"),
+    State("reports-month-select", "value"),
+    prevent_initial_call=True,
+)
+def _download_reports_pdf(_n_clicks, month_value):
+    ctx_data = _build_report_ctx()
+    ctx_data["monthly_notes"] = monthly_report_notes.NOTES.get(month_value)
+    pdf_bytes = report.build_report_pdf(ctx_data, month_value)
+    label = next((o["label"] for o in _REPORTS_MONTH_OPTIONS if o["value"] == month_value), month_value)
+    filename = f"TXproducts_LinkedIn_Report_{label.replace(' ', '_')}.pdf"
+    return dcc.send_bytes(lambda buf: buf.write(pdf_bytes), filename=filename)
+
+
 if __name__ == "__main__":
     app.run(debug=False, port=8051)
diff --git a/build_static_export.py b/build_static_export.py
index 20ffc49..e639a09 100644
--- a/build_static_export.py
+++ b/build_static_export.py
@@ -118,6 +118,15 @@ def render(comp):
     cls = type(comp).__name__
     module = type(comp).__module__
 
+    # --- Reports tab content area: empty in the live layout (filled by a callback), so
+    # bake in the one selectable month's rendered output directly. Switching months
+    # needs the live tool (see the disabled dropdown below), same rationale as every
+    # other live-only control in this exporter.
+    if getattr(comp, "id", None) == "reports-content-area":
+        if A._REPORTS_DEFAULT_MONTH:
+            return render(A._render_reports_tab(A._REPORTS_DEFAULT_MONTH))
+        return render(dbc.Alert("No campaign months have been written up yet.", color="info"))
+
     # --- Plotly graphs ---
     if isinstance(comp, dcc.Graph):
         return register_chart(comp.figure, getattr(comp, "style", None), getattr(comp, "id", None))
@@ -131,10 +140,13 @@ def render(comp):
         style = style_dict_to_css(getattr(comp, "style", None))
         extra_class = getattr(comp, "className", "") or ""
         cid = getattr(comp, "id", "")
-        if cid == "report-timeframe-select":
+        if cid in ("report-timeframe-select", "reports-month-select"):
+            title = ("Switching months needs the live tool, not available in this snapshot."
+                      if cid == "reports-month-select" else
+                      "PDF generation needs the live tool, not available in this snapshot.")
             return (f'<select id="{esc(cid)}" class="form-select compact-dropdown {esc(extra_class)}" '
                     f'style="{style};opacity:0.5;cursor:not-allowed" disabled '
-                    f'title="PDF generation needs the live tool, not available in this snapshot.">{opts}</select>')
+                    f'title="{title}">{opts}</select>')
         return f'<select id="{esc(cid)}" class="form-select compact-dropdown {esc(extra_class)}" style="{style}" onchange="window.__onSelect && window.__onSelect(this)">{opts}</select>'
 
     # --- Stores/Download: no static equivalent needed (Download requires a live backend) ---
@@ -184,6 +196,15 @@ def render(comp):
             md = getattr(comp, "md", None)
             colcls = f"col-md-{md}" if md else "col"
             return f'<div class="{colcls} {esc(extra_class)}" style="{style}"{id_attr}>{children_html}</div>'
+        if cls == "Table":
+            table_classes = "table"
+            if getattr(comp, "bordered", False):
+                table_classes += " table-bordered"
+            if getattr(comp, "striped", False):
+                table_classes += " table-striped"
+            if getattr(comp, "size", None):
+                table_classes += f" table-{comp.size}"
+            return f'<table class="{table_classes} {esc(extra_class)}" style="{style}"{id_attr}>{children_html}</table>'
         if cls == "Card":
             return f'<div class="card {esc(extra_class)}" style="{style}"{id_attr}>{children_html}</div>'
         if cls == "CardHeader":
@@ -236,7 +257,7 @@ def render(comp):
         if cls == "Hr":
             return f'<hr style="{style}"{class_attr}/>'
         if cls == "Button":
-            if cid == "generate-report-btn":
+            if cid in ("generate-report-btn", "reports-download-btn"):
                 disabled_style = style_dict_to_css({**{k: v for k, v in (getattr(comp, "style", None) or {}).items()},
                                                      "opacity": "0.5", "cursor": "not-allowed"})
                 return (f'<button type="button"{id_attr}{class_attr} style="{disabled_style}" disabled '
diff --git a/follower_directory_data.py b/follower_directory_data.py
new file mode 100644
index 0000000..693bf6e
--- /dev/null
+++ b/follower_directory_data.py
@@ -0,0 +1,49 @@
+"""Reads audience composition straight from assets/follower_directory.html's DATA array,
+so the dashboard and PDF reports show the same numbers as the hand-maintained Follower
+Study/Directory instead of a second, independently hand-computed set that can drift out
+of sync with it.
+
+Deliberately has zero import-time dependency on app.py, same reasoning as report.py's
+module docstring: app.py runs as "__main__", so importing it here would re-execute it as
+a second, still-mid-import copy.
+"""
+import json
+import os
+import re
+from collections import Counter
+
+
+def load_directory(assets_dir):
+    """Parses the `const DATA = [...];` array out of follower_directory.html into a list
+    of dicts. Returns an empty list if the file is missing rather than raising, since a
+    fresh checkout or a stripped-down deployment may not have it yet."""
+    path = os.path.join(assets_dir, "follower_directory.html")
+    if not os.path.exists(path):
+        return []
+    text = open(path, encoding="utf-8").read()
+    match = re.search(r"const DATA = (\[.*?\]);", text, re.S)
+    if not match:
+        return []
+    return json.loads(match.group(1))
+
+
+def composition_stats(data, month_value=None, top_countries=None):
+    """Sector/seniority/function/country breakdowns, either over the whole audited set
+    (month_value=None) or restricted to the entries whose `followedSort` matches
+    month_value (e.g. "2026-07"), i.e. that month's own follower cohort as profiled so
+    far. A month's count only ever grows as later audit passes catch up, it is a
+    coverage figure, not a final one.
+
+    top_countries defaults to unbounded (None = every country shown): this was
+    previously hardcoded to 10, which silently dropped tail countries from a monthly
+    cohort's breakdown while the "total" count kept counting them, the table's own sum
+    fell short of the total it was supposed to explain (caught when adding Sabine E.
+    Herlitschka's July correction pushed the cohort to 11 distinct countries)."""
+    rows = data if month_value is None else [d for d in data if d.get("followedSort") == month_value]
+    return {
+        "total": len(rows),
+        "sector": Counter(d.get("sector", "Unknown") for d in rows),
+        "seniority": Counter(d.get("seniority", "Unknown") for d in rows),
+        "function": Counter(d.get("function", "Unknown") for d in rows),
+        "country": Counter(d.get("country") or "Unknown/blank" for d in rows).most_common(top_countries),
+    }
diff --git a/monthly_report_notes.py b/monthly_report_notes.py
new file mode 100644
index 0000000..95c96fe
--- /dev/null
+++ b/monthly_report_notes.py
@@ -0,0 +1,66 @@
+"""Curated, hand-written narrative for each month of the 3-month LinkedIn campaign
+(July, August, September 2026, plus an end-of-campaign report). Growth numbers are
+computed live from the exports elsewhere (see report.compute_period_stats), this file
+only holds the "why", which lives in git history and project context, not in any
+dataframe. Keyed by the same "YYYY-MM" strings report.report_month_options() produces.
+
+Add August's and September's entries here as those months are written up, no other
+code changes are needed for the Reports tab or PDF to pick them up.
+"""
+
+NOTES = {
+    "2026-07": {
+        "features_shipped": [
+            "Initial dashboard build (Jul 21): Overview, Content, Followers, Visitors, "
+            "Competitors, and Growth tabs, weekly-pace tracking, PDF report generation, "
+            "data upload with live refresh, and a static HTML export builder for sharing "
+            "via NextCloud.",
+            "Posts table made scrollable instead of growing indefinitely, and the file "
+            "finder hardened to ignore non-export files sitting in the same folder (Jul 22).",
+            "Follower Directory and Follower Study linked directly from the Followers tab "
+            "hero banner (Jul 31).",
+            "Followers by Country map expanded from 14 hardcoded country centroids to a "
+            "near-complete world table, with a console warning for any country still "
+            "missing one (Jul 31).",
+            "Calendar month drill-down tiles, a follower-breakdown expand toggle, a CTR "
+            "explainer, and manual post title overrides added across the Followers and "
+            "Content tabs (Jul 31).",
+        ],
+        "data_quality": [
+            "File finder no longer crashes when a stray non-export file, a manual roster "
+            "or an Office lock file, sits in the same folder as the real LinkedIn exports.",
+            "Country map no longer silently drops a country if it is missing a centroid, "
+            "it now warns in the console instead, and the on-screen country count is "
+            "computed from the same rollup as the map so the two can never disagree.",
+            "Visitors category breakdown (Location, Seniority, Company size) is no longer "
+            "lost month to month. LinkedIn only reports a category once it clears a "
+            "per-export minimum-visits threshold, so a single-snapshot approach was "
+            "silently losing real categories that dipped below the threshold one week and "
+            "reappeared the next. Replaced with an append-only running-total log "
+            "(visitor_history.py).",
+            "Conversion Rate chart no longer silently drops days with zero tracked "
+            "visitors but a real follower change (LinkedIn lets people follow a Page "
+            "directly from a feed post, with no page visit at all). Every day now shows, "
+            "with those days flagged instead of hidden.",
+        ],
+        "follower_study": [
+            "Grew and was actively reclassified throughout July as new followers were "
+            "screenshotted and profiled by hand.",
+            "A seniority and sector classification fix landed Jul 31 (backup preserved as "
+            "follower_directory_BACKUP-before-student-sector-fix.html), plus several "
+            "individual profile confirmations (Dominique Elser, Anna Bujak, Emma Xia, "
+            "Marten Edwards, Mudabar Ahmad Mubashar, Martin Schroter) that corrected "
+            "taglines into real titles and employers.",
+            "Not yet brought under git version control during July itself, that happened "
+            "Aug 3 and falls outside this report.",
+        ],
+        "outstanding": [
+            "The end-of-week report explaining every dashboard chart to the bosses was "
+            "due Fri Jul 31, status needs confirming with the user before marking this "
+            "done.",
+            "The conversion-rate weekly-bucket and 3 to 7 day match-window widening "
+            "discussed Jul 28, to make the follow-attribution pattern more convincing "
+            "across multiple weeks, is still conceptual, not implemented.",
+        ],
+    },
+}
diff --git a/report.py b/report.py
index c0ba968..b43b9d8 100644
--- a/report.py
+++ b/report.py
@@ -20,6 +20,8 @@ from reportlab.lib.units import cm
 from reportlab.platypus import (SimpleDocTemplate, Table, TableStyle, Paragraph,
                                  Spacer, Image, HRFlowable)
 
+import follower_directory_data
+
 NAVY = colors.HexColor("#1b3a5c")
 LIGHT_GRAY = colors.HexColor("#f0f4f8")
 TEXT_GRAY = colors.HexColor("#6b7a8d")
@@ -62,19 +64,12 @@ def _fig_to_image(fig, width_cm=17, height_px=300, width_px=900):
     return Image(buf, width=width_cm * cm, height=width_cm * cm * aspect)
 
 
-def build_report_pdf(ctx, timeframe_value="all"):
-    """ctx is a dict bundling everything the report needs from the live app:
-    D, posts, comp_ordered, current_total_followers, daily_followers_figure (the
-    figure-builder function), std_fn (the _std styling helper), TMPL, FORMAT_COLORS."""
+def compute_period_stats(ctx, timeframe_value="all"):
+    """Filters D/posts down to timeframe_value and computes the headline numbers shared
+    by the PDF report and the live Reports tab, so both always show the same figures
+    instead of two independently-computed versions drifting apart."""
     D = ctx["D"]
     posts = ctx["posts"]
-    comp_ordered = ctx["comp_ordered"]
-    current_total_followers = ctx["current_total_followers"]
-    daily_followers_figure = ctx["daily_followers_figure"]
-    std_fn = ctx["std_fn"]
-    tmpl = ctx["TMPL"]
-    format_colors = ctx["FORMAT_COLORS"]
-    growth_snapshot = ctx["growth_snapshot"]
 
     fts_f = _filtered(D["fts"], "Date", timeframe_value)
     posts_f = _filtered(posts, "Created date", timeframe_value)
@@ -87,7 +82,84 @@ def build_report_pdf(ctx, timeframe_value="all"):
     posts_published = len(posts_f)
     best_eng = float(posts_f["Eng %"].max()) if len(posts_f) else 0.0
     avg_eng_series = cts_f["Engagement rate (total)"].replace(0, float("nan")) if not cts_f.empty else pd.Series(dtype=float)
-    avg_eng = float(avg_eng_series.mean()) if not avg_eng_series.dropna().empty else 0.0
+    # "Engagement rate (total)" is stored as a 0-1 fraction (confirmed against raw export
+    # data, max observed 5.9 i.e. 590%), not already a percentage like "Eng %"/"CTR %" —
+    # scale to percentage-point form so the "%" already in every display string is correct.
+    avg_eng = float(avg_eng_series.mean()) * 100 if not avg_eng_series.dropna().empty else 0.0
+
+    # Monthly Growth Rate: this month's net new followers as a percentage of the
+    # cumulative total the page already had at the start of that month, resets every
+    # month against its own starting count rather than a single fixed campaign baseline
+    # (e.g. July: 22 new / 85 at start of July = +25.9%). Only defined for a specific
+    # "YYYY-MM" timeframe_value, not "all" or a whole year, since "start of the period"
+    # is only a clean concept for a single calendar month.
+    start_of_month_followers = None
+    monthly_growth_rate = None
+    if timeframe_value and timeframe_value != "all" and not timeframe_value.startswith("year-"):
+        year, month = (int(p) for p in timeframe_value.split("-"))
+        first_day = pd.Timestamp(year=year, month=month, day=1)
+        prior = D["fts"][D["fts"]["Date"] < first_day]
+        if not prior.empty:
+            start_of_month_followers = int(prior["Cumulative Followers"].iloc[-1])
+            if start_of_month_followers:
+                monthly_growth_rate = (new_followers / start_of_month_followers) * 100
+
+    return {
+        "fts_f": fts_f, "posts_f": posts_f, "cts_f": cts_f, "vts_f": vts_f,
+        "new_followers": new_followers, "total_impressions": total_impressions,
+        "total_page_views": total_page_views, "posts_published": posts_published,
+        "best_eng": best_eng, "avg_eng": avg_eng,
+        "start_of_month_followers": start_of_month_followers,
+        "monthly_growth_rate": monthly_growth_rate,
+        "period_label": _period_label(timeframe_value, D["fts"]),
+    }
+
+
+def _bullets(items, style):
+    return [Paragraph(f"&bull;&nbsp; {item}", style) for item in items]
+
+
+def _counter_rows(counter, header_left, top_n=None):
+    """A Counter (or list of (key, count) pairs, for the `country` field
+    composition_stats() returns) as reportlab Table rows, header included.
+    No truncation by default, sector/seniority/function are a small fixed taxonomy and
+    country is uncapped by composition_stats() for the same reason, so the table's own
+    sum matches the cohort total instead of silently dropping the tail categories."""
+    items = counter.most_common(top_n) if hasattr(counter, "most_common") else list(counter)[:top_n]
+    return [[header_left, "Count"]] + [[str(k), str(v)] for k, v in items]
+
+
+def build_report_pdf(ctx, timeframe_value="all"):
+    """ctx is a dict bundling everything the report needs from the live app:
+    D, posts, comp_ordered, current_total_followers, daily_followers_figure (the
+    figure-builder function), std_fn (the _std styling helper), TMPL, FORMAT_COLORS,
+    growth_snapshot, and optionally monthly_notes (a curated per-month narrative dict,
+    see monthly_report_notes.py) and directory_data (the parsed Follower Directory,
+    see follower_directory_data.py). monthly_notes/directory_data are only rendered
+    when monthly_notes is present, so "All Time"/year-level reports are unaffected."""
+    D = ctx["D"]
+    posts = ctx["posts"]
+    comp_ordered = ctx["comp_ordered"]
+    current_total_followers = ctx["current_total_followers"]
+    daily_followers_figure = ctx["daily_followers_figure"]
+    std_fn = ctx["std_fn"]
+    tmpl = ctx["TMPL"]
+    format_colors = ctx["FORMAT_COLORS"]
+    growth_snapshot = ctx["growth_snapshot"]
+    monthly_notes = ctx.get("monthly_notes")
+    directory_data = ctx.get("directory_data") or []
+
+    period = compute_period_stats(ctx, timeframe_value)
+    fts_f, posts_f, cts_f = period["fts_f"], period["posts_f"], period["cts_f"]
+    new_followers = period["new_followers"]
+    total_impressions = period["total_impressions"]
+    total_page_views = period["total_page_views"]
+    posts_published = period["posts_published"]
+    best_eng = period["best_eng"]
+    avg_eng = period["avg_eng"]
+    monthly_growth_rate = period["monthly_growth_rate"]
+    start_of_month_followers = period["start_of_month_followers"]
+    period_label = period["period_label"]
 
     styles = getSampleStyleSheet()
     title_style = ParagraphStyle("TXTitle", parent=styles["Title"], textColor=NAVY, fontSize=20, spaceAfter=2)
@@ -96,8 +168,6 @@ def build_report_pdf(ctx, timeframe_value="all"):
     body_style = ParagraphStyle("TXBody", parent=styles["Normal"], fontSize=9, textColor=colors.HexColor("#333333"))
     note_style = ParagraphStyle("TXNote", parent=styles["Normal"], fontSize=8, textColor=TEXT_GRAY, spaceBefore=4)
 
-    period_label = _period_label(timeframe_value, D["fts"])
-
     story = []
     story.append(Paragraph("TXproducts, LinkedIn Performance Report", title_style))
     story.append(Paragraph(
@@ -110,12 +180,13 @@ def build_report_pdf(ctx, timeframe_value="all"):
     story.append(Paragraph("Key Metrics", h2_style))
     kpi_data = [
         ["Total Followers\n(as of today)", "New Followers", "Posts Published", "Total Impressions",
-         "Total Page Views", "Best Post Eng.", "Avg. Eng. Rate"],
+         "Total Page Views", "Best Post Eng.", "Avg. Eng. Rate", "Monthly Growth Rate"],
         [f"{current_total_followers:,}", f"{new_followers:,}", f"{posts_published:,}",
          f"{total_impressions:,}", f"{total_page_views:,}", f"{best_eng:.1f}%",
-         f"{avg_eng:.2f}%" if avg_eng else "—"],
+         f"{avg_eng:.2f}%" if avg_eng else "—",
+         f"{monthly_growth_rate:+.1f}%" if monthly_growth_rate is not None else "—"],
     ]
-    kpi_table = Table(kpi_data, colWidths=[2.6 * cm] * 7)
+    kpi_table = Table(kpi_data, colWidths=[2.25 * cm] * 8)
     kpi_table.setStyle(TableStyle([
         ("BACKGROUND", (0, 0), (-1, 0), LIGHT_GRAY),
         ("TEXTCOLOR", (0, 0), (-1, 0), TEXT_GRAY),
@@ -130,6 +201,14 @@ def build_report_pdf(ctx, timeframe_value="all"):
         ("BOTTOMPADDING", (0, 0), (-1, -1), 6),
     ]))
     story.append(kpi_table)
+    if monthly_growth_rate is not None:
+        story.append(Paragraph(
+            f"Monthly Growth Rate: this month's net new followers as a percentage of the "
+            f"followers already on the page at the start of the month. {period_label} started "
+            f"at {start_of_month_followers:,} and gained {new_followers:,}, "
+            f"<b>{monthly_growth_rate:+.1f}%</b>. Resets every month against that month's own "
+            f"starting count, not a single fixed baseline for the whole campaign.",
+            note_style))
 
     # --- Weekly growth pace ---
     story.append(Paragraph("Weekly Growth Pace", h2_style))
@@ -172,6 +251,52 @@ def build_report_pdf(ctx, timeframe_value="all"):
         f"unreachable target permanently.",
         note_style))
 
+    # --- Monthly narrative sections (only when curated notes exist for this period) ---
+    if monthly_notes:
+        story.append(Paragraph("Features / Fixes Shipped", h2_style))
+        story.extend(_bullets(monthly_notes.get("features_shipped", []), body_style))
+
+        story.append(Paragraph("Data Quality Issues Found and Fixed", h2_style))
+        story.extend(_bullets(monthly_notes.get("data_quality", []), body_style))
+
+        story.append(Paragraph("Audience Composition (Follower Directory Coverage)", h2_style))
+        cohort = follower_directory_data.composition_stats(directory_data, timeframe_value)
+        all_time = follower_directory_data.composition_stats(directory_data)
+        if cohort["total"]:
+            story.append(Paragraph(
+                f"LinkedIn recorded {new_followers} new followers in {period_label}. Of those, "
+                f"<b>{cohort['total']} have been identified and profiled</b> in the Follower "
+                f"Directory so far (a separate, hand-maintained audit of who our followers are, "
+                f"not a second follower count), out of {all_time['total']} audited across the "
+                f"whole page to date.", body_style))
+            for field, label in (("sector", "Sector (this month)"),
+                                 ("seniority", "Seniority (this month)"),
+                                 ("country", "Country (this month)")):
+                comp_table = Table(_counter_rows(cohort[field], label), colWidths=[9 * cm, 3 * cm])
+                comp_table.setStyle(TableStyle([
+                    ("BACKGROUND", (0, 0), (-1, 0), LIGHT_GRAY),
+                    ("FONTSIZE", (0, 0), (-1, -1), 8),
+                    ("GRID", (0, 0), (-1, -1), 0.5, colors.HexColor("#e0e7ef")),
+                    ("TOPPADDING", (0, 0), (-1, -1), 4), ("BOTTOMPADDING", (0, 0), (-1, -1), 4),
+                ]))
+                story.append(comp_table)
+                story.append(Spacer(1, 6))
+        else:
+            story.append(Paragraph(
+                f"LinkedIn recorded {new_followers} new followers in {period_label}, but none "
+                f"have been identified and profiled in the Follower Directory yet.", body_style))
+        story.append(Paragraph(
+            "The Follower Directory is a hand-maintained audit that lags real LinkedIn "
+            "follows and gets corrected after the fact as more profiles are screenshotted "
+            "and cross-referenced, so this is a coverage figure that can still grow in "
+            "later updates, not a final count.", note_style))
+
+        story.append(Paragraph("Follower Study / Directory Updates", h2_style))
+        story.extend(_bullets(monthly_notes.get("follower_study", []), body_style))
+
+        story.append(Paragraph("Outstanding", h2_style))
+        story.extend(_bullets(monthly_notes.get("outstanding", []), body_style))
+
     # --- Follower growth chart ---
     if not fts_f.empty:
         story.append(Paragraph("Follower Growth", h2_style))
@@ -179,12 +304,18 @@ def build_report_pdf(ctx, timeframe_value="all"):
         fig.update_layout(showlegend=True, font=dict(size=14))
         story.append(_fig_to_image(fig))
 
-    # --- Top posts table ---
-    story.append(Paragraph("Top Posts" + ("" if timeframe_value == "all" else f", {period_label}"), h2_style))
+    # --- Posts table: every post for a monthly campaign report (monthly_notes present),
+    # otherwise the top 5 by engagement, same as before, for the generic sidebar PDF
+    # control where "All Time"/a whole year could mean far too many rows to list in full.
+    if monthly_notes:
+        story.append(Paragraph(f"All Posts, {period_label}", h2_style))
+        shown_posts = posts_f.sort_values("Created date")
+    else:
+        story.append(Paragraph("Top Posts" + ("" if timeframe_value == "all" else f", {period_label}"), h2_style))
+        shown_posts = posts_f.sort_values("Eng %", ascending=False).head(5)
     if len(posts_f):
-        top_posts = posts_f.sort_values("Eng %", ascending=False).head(5)
         rows = [["Post", "Date", "Type", "Impressions", "Eng %", "CTR %"]]
-        for _, r in top_posts.iterrows():
+        for _, r in shown_posts.iterrows():
             rows.append([
                 Paragraph(str(r["Short title"])[:60], body_style),
                 r["Created date"].strftime("%d/%m/%Y"),
diff --git a/requirements.txt b/requirements.txt
index 6b222c3..75cc868 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -4,3 +4,5 @@ plotly>=5.0
 pandas>=1.3
 openpyxl>=3.0
 xlrd>=2.0
+reportlab>=4.0
+kaleido>=0.2