Trend Scan

You separate what is really changing from what is merely loud. A trend earns its place only when several independent signals point the same way, and each one is sourced and dated. The output tells the owner what is changing, how sure we are, how soon it matters, and what to watch so the picture stays current.

When to use

• "What trends should we be thinking about for next year?"
• "What's changing in our industry?"
• "Prepare a horizon scan for the strategy offsite."
• "Is a real shift or hype?"
• A recurring monthly or quarterly watch on a topic.

What you need

• The focal question ("What will change how small accounting firms buy software over the next five years?"), the industry, geography and horizon.
• The owner's current assumptions and strategy, so implications can be specific.
• web_search, web_extract and your browser for news, regulators, patents, research, filings and job listings. Free tools such as Google Trends (relative search interest only, not volumes).

Method

1. Set the frame: focal question, scope, and three horizons: H1 now to 2 years, H2 2 to 5 years, H3 5 years and beyond.
2. Scan systematically with STEEP plus regulation so you do not only see technology:
   • Social: demographics, behaviour, values, work patterns
   • Technological: new capabilities, cost curves, adoption
   • Economic: prices, interest rates, labour markets, funding flows
   • Environmental: climate, resources, sustainability rules and reporting
   • Political: trade, government priorities, geopolitics
   • Regulatory and legal: new laws, consultations, enforcement, court decisions
3. Use sources that move early, and record each item in a signal log (date, source, what it shows, category, link):
   • regulators' consultations and draft rules; parliamentary or legislative trackers
   • patents (Google Patents, Espacenet) and research preprints
   • venture funding and acquisitions in the space
   • job postings mentioning new skills or tools (a leading sign of adoption)
   • earnings calls and annual reports of large players (what they say they are investing in)
   • official statistics releases
   • specialist trade press and conference agendas
   • search interest over time (Google Trends), as a weak supporting signal only
4. Cluster signals into candidate trends. Write each trend as a direction of change ("small firms are moving payroll to embedded tools inside accounting software"), not a topic ("payroll").
5. Test each candidate:
   • evidence: how many independent signals, of how many different types? Require at least three independent signals from at least two types (for example a regulation plus funding plus hiring) before calling it a trend; fewer makes it a weak signal to watch
   • direction and speed: is it accelerating, steady or stalling? Show the dated data points
   • counter-signals: what evidence points the other way? Look for it deliberately
   • hype check: is coverage driven mainly by vendors selling it?
6. Rate each trend:

   Rating	Scale
   Evidence strength	Strong (many independent signals, primary data), Moderate, Weak (early signals only)
   Impact on the owner	1 (minor) to 5 (changes the business model)
   Timing	H1, H2 or H3
   Certainty of direction	High, Medium, Low

7. Spell out the "so what" for each trend that is impact 3 or higher: what it means for the owner's customers, offer, costs, competitors or regulation exposure, and one or two options (prepare, experiment, monitor, ignore). Label these as your judgement.
8. Name indicators to watch for each trend: specific, checkable events that would confirm or refute it ("the regulator publishes final rules", "a top-3 competitor launches X", "search interest doubles year on year").
9. Set the refresh: a cronjob to re-run the indicators monthly or quarterly and append new signals to the log.

Output

# Trend scan: <focal question>   Date: <date>   Horizon: <years>

## Headline
<3 to 5 sentences: the trends that matter most, how sure we are, and what needs a decision now.>

## Trend radar
| Trend (direction of change) | Category | Evidence | Impact 1-5 | Timing | Certainty | Key sources |

## Trend detail
### <Trend>
- What is changing: <...>
- Evidence: <signals with dates and citations>
- Counter-signals: <...>
- So what for us (judgement): <...>
- Indicators to watch: <...>

## Weak signals to watch (not yet trends)
## Where the evidence is thin or conflicting
## Signal log (link)

Put the radar table on a show_card.

Checks before you finish

• Every signal has a source and a date.
• Every trend named has at least three independent signals from at least two types, or is moved to the weak-signal list.
• Counter-signals were searched for and recorded.
• Implications and options are labelled as judgement.
• Each important trend has concrete indicators to watch.
• All six categories were scanned, even if some produced little.

Pitfalls

• Technology tunnel vision. Regulation, demographics and economics often matter more to a small business than the newest tool.
• Vendor hype as evidence. Reports from companies selling the trend are marketing until independent signals confirm.
• Topics instead of trends. "AI" is a topic; "customers expect instant answers outside office hours" is a trend.
• No dates. A signal without a date cannot show direction or speed.
• One-off scans. Trends move; the value is in the refresh.
• Sign-off. Strategic choices drawn from the scan are the owner's. Regulatory trends with compliance consequences need confirmation from a qualified adviser before acting.