What we deliver

One programme. Several outputs, chosen to the question.

Everything below comes out of the same programme: a baseline read of the public conversation, then tracking on the rhythm your team chooses. The outputs differ because the questions do. Open any of them.

Question 3 of 8 · March
Which side effects do patients discuss before starting, and has that changed since the baseline?
Fatigue has overtaken nausea as the side effect patients raise before their first cycle. The shift began in the week of the congress read-out and has held for five weeks. Nausea talk is unchanged in volume but now comes mostly from patients already on treatment.
▲ Fatigue pre-treatment mentions +38% vs baseline
Pre-treatment side-effect mentions, share of patient posts
Fatigue
41%
Nausea
32%
Hair loss
17%
Other
10%
n = 212 patient and caregiver posts · baseline: fatigue 29%
PATIENT · forum
"Starting next week and honestly the tiredness is what scares me more than feeling sick, everyone on here says it just flattens you."
Source link in the internal appendix
CAREGIVER · Reddit
"Mum's oncologist mentioned fatigue almost in passing. We had no idea it would mean whole days in bed."
Source link in the internal appendix
PATIENT · Instagram
"Cycle 4 done. Nausea is manageable with the meds now, it's the exhaustion nobody warned me about."
Source link in the internal appendix
Patient and caregiver posts only · clinicians, news and brand-owned excludedSTARLING · SAMPLE · INVENTED DATA

The answer report

Monthly or quarterly. The questions your team agreed at the start, answered again against the baseline. One question per page.

  • 1
    Your question, in your words
    The report is built on a bank of eight to twelve questions agreed at scoping, each tagged as something we count, something we read, or something only your stakeholders can answer. The question is the page title. There is no dashboard to interpret.
  • 2
    The answer in one paragraph
    Written by an analyst who read the posts, not generated from a chart. It says what is happening, when it started, and what it is not.
  • 3
    The count behind it
    Every figure names its base: how many posts, which audience. Counts come from the platform; the classification of who is speaking is ours.
  • 4
    What changed since the baseline
    Each answer is a delta against the landscape study, then against last month, so the report reads as movement rather than a snapshot.
  • 5
    The voices behind the number
    Three to five verbatims per answer, each labelled by who is speaking and where. Patients and caregivers are separated from clinicians, news and brand-owned accounts by our classifier, then checked by a person wherever a post carries weight.
  • 6
    Privacy, by design
    Private individuals are never named. Quotes are lightly paraphrased. Source links go into a separate internal appendix under the agreed pharmacovigilance and privacy protocol, not the circulated report.

Every page of the report

Click a spread
Answer report · March
Cover
Cover
At a glance
1 page
The month in one page
The month in one paragraph
1 page
Narrative summary
Q1 to Q8, one per page
8 pages
The question bank, answered
Weekly volume
1 page
Volume and velocity
Channels
1 page
Where the conversation lives
Congress week
Timeline
Event timeline
Category context
Share of voice
Comparator share of voice
In their words
2 pages
Verbatims
Next month
1 page
What we will watch
Appendix A · Adverse events
Appendix
Adverse-event handling
Appendix B · Method
Appendix
Method and sources
The question bank, answered
Eight to twelve pages, one question each, in the format shown above. The bank is agreed at scoping and can change at any quarter; a question that stops mattering is retired and its page goes.
Pages 4 to 11

The rest of the family

Same programme, same classification, same team. Each output exists because a different kind of question needs a different shape.

Landscape study
Once · the baseline
Landscape study
Twelve months of public conversation read once, so every later report has something to measure against.
  • The question bank answered for the first time
  • Community map: where each audience talks
  • Vernacular glossary: how patients name the drug, the test, the side effects
  • Education gaps and misinformation appendix
  • Influencer map: organisations and public figures only
Share of voice
Monthly · the counting layer
Share of voice
Who is being talked about, and who AI assistants talk about, month by month.
  • Social: brands and products by channel, originals and reposts separated
  • AI: which brands and products ChatGPT, Gemini and Google's AI answers name when a patient, clinician or payer asks, and which sources they cite
  • Trend against the baseline; a paragraph on what moved
Alerts and event briefs
When something happens
Alerts and event briefs
A congress, an approval, a read-out, a competitor move. A short read soon after it happens, not at month end.
  • What was said, by whom, how it spread
  • Patient reaction separated from professional and press
  • What it changes for the next report
Data exports and factbook pages
Monthly · for your own analysts
Data exports and factbook pages
The classified data behind every report, in a workbook your team can cut, and pages for your annual brand factbook.
  • Every post with audience, theme and channel tags
  • Master trend sheet, month over month
  • Factbook contributions in your template
Workshops and capability building
On request
Workshops and capability building
Turn a report into decisions, or build the skill in your own team.
  • 90-minute insight workshop with a one-page decision record
  • Social listening tool training, taxonomy design, AI in the CI workflow
  • Half-day, full-day or multi-session
Intelligence infrastructure
For teams running their own programme
Intelligence infrastructure
The pipeline behind our reports, designed and run for your analysts: collection, AI classification, review queue, adverse-event screening.
  • Your analysts read the fifty posts that matter, not the five thousand
  • Built on the taxonomies we have refined since 2012

How they fit together

A programme runs in this order. The rhythm of the last step is your choice, and year two starts there without a new landscape.

Step 1
Scope
Agree the question bank, the audiences, the comparator set and the adverse-event protocol with your pharmacovigilance team.
Question bankAE protocol
Step 2
Landscape
Twelve months read once. The baseline every later report measures against.
Landscape studyShare of voiceData export
Step 3
Calibrate
A session with the brand team: which answers surprised, which questions to retire, what to watch.
Workshop
Step 4 · ongoing
Track
Monthly or quarterly. Continuous collection and adverse-event screening either way.
Answer reportShare of voiceEvent briefsData export

What sits behind every output

Public sources only; no closed groups. Audiences separated by classification, then checked by a person. Adverse events screened continuously and handled to your protocol, on business days, by trained readers. No private individual is ever named. The full list of what we do and do not do is on each door page.

See one for your brand

Tell us the brand or therapy area and what decision the work would feed. We will come back with what the public conversation looks like and which outputs fit.

Get in touch