Campaign ideas from your market, surveys and data.
Your campaign earns coverage when the idea rests on a finding nobody else has. Fractl Agents helps you find those ideas, test them and back them with original data. You get there sooner, with numbers an editor can check.
Give me campaign ideas about coffee that a journalist would cover.
Ranked by strength
- 1The real cost of a daily coffee habit, by cityData storyOriginal survey plus public price data; a local version for 20 cities
- 2How stale is the coffee on supermarket shelves?InvestigationBuy, test, date; a finding nobody has published
- 3The office coffee indexSurvey1,000 office workers on what they drink and what they tolerate
Limitation: strength is a judgement from past coverage patterns, not a prediction.
Campaign ideas
A pitch lands when the idea is new and provable. From a bare topic to an idea scored and checked against past coverage, Fractl Agents does the groundwork and you pick the winner.
Generate campaign ideas from the patterns in campaigns that won coverage in other industries, scored and ranked by strength.
Stress-test an idea against prior coverage and live search data, and get a score, the weak spots ranked and a fix for each.
Compare what the top-ranking pages say on a topic with what experts, forums and new research discuss, and get the missed angles as briefs.
Turn a brand’s website into story ideas for each city or region the brand serves, anchored on a public statistic about that place.
Spot topics growing fast across search, social, news and video before they peak, and get campaign angles to match.
Find calculator and quiz ideas ranked by search demand and by how many sites link to similar ones, each with a build spec.
Give me campaign ideas about coffee that a journalist would cover.
For this question I’ll go with ideas built from what has worked before. Other ways I could go:
- Ideas built from what has worked beforePatterns from covered campaigns in other subjectsChosen
- Check an idea before we commitAlready common, already covered, hard to substantiate?
- What the top articles missCompared with what experts and practitioners say
- Content topics worth owningSearch demand, audience questions, your own advantages
- Local angles and timely commentaryTied to a place, or to this week’s news
- A calculator or quiz conceptSomething useful people would share
Some of the ways. Say the word and I’ll take another.
Ideas it is. I’ll look at what got coverage in other subjects and bring the patterns to coffee.
Ranked by strength
- 1The real cost of a daily coffee habit, by cityData storyOriginal survey plus public price data; a local version for 20 cities
- 2How stale is the coffee on supermarket shelves?InvestigationBuy, test, date; a finding nobody has published
- 3The office coffee indexSurvey1,000 office workers on what they drink and what they tolerate
- 4Which coffee gets the most complaints?ReviewsMarket Voice across 40 brands; a league table
Limitation: strength is a judgement from past coverage patterns, not a prediction.
Idea 1 it is. Here is the research plan:
- 1Survey 1,000 coffee drinkersSpend, habit, city
- 2Pull public price data for 20 citiesCoffee, milk, rent as the comparison
- 3Write the findings and a city tableOne national story, 20 local versions
Campaign ideas ranked by strength, then a research plan for the chosen idea.
Original surveys
A survey gives you numbers nobody else has, if the questions hold up. Fractl Agents checks every question with you before a single response comes in, then reads the results.
Design a survey built for a news story: question areas first, then every question, the screening and the follow-up logic, for your review.
Check a questionnaire you’ve written for leading wording, bias, question order and broken follow-up logic, question by question.
Build the approved survey as a draft in your connected SurveyMonkey account, ready for you to send.
Analyze the completed responses, test which differences between groups hold up statistically, and sort every open-text answer into themes.
Get the findings sorted into newsworthy and obvious, pick the ones to lead with, and receive tables, charts and a written report.
Design a survey about how people really make coffee at home.
For this question I’ll go with design the survey. Other ways I could go:
- Design the surveyQuestion areas first, then every question for reviewChosen
- Review a questionnaire we wroteWording, bias, order, follow-up logic
- Analyze the responsesClean the data, compare groups, find the themes
- Draft it in SurveyMonkeyIn a connected account
- Turn findings into a reportTables, charts and a write-up
- Improve the next questionnaireFrom what the last one taught you
Some of the ways. Say the word and I’ll take another.
I’ll propose the question areas first, then every question, for your review.
- Brewing method and equipment
- Money spent on coffee at home
- Where the beans come from
- Ritual, mood and the first cup
- How much Americans spend on coffeeAlready in public data
Apply Fractl Agency Survey Rules
- Single choiceHow do you usually make coffee at home?Looks good
- Multiple choiceWhich of these have you bought in the last year?Looks good
- NumberAbout how much do you spend on coffee at home in a month?Looks good
- ScaleHow much do you agree: “I know where my coffee comes from.”Looks good
- 14 more questions
Approved and built in your SurveyMonkey account. The 1,000 responses are due in four days.
A survey plan, reviewed question by question before the draft is built.
Your own data
Your own files and public data often hide a story nobody has told. Even one messy export is enough for Fractl Agents to help you find that story and check every number.
Mine an export of thousands of rows for story angles, from rankings and trends to correlations, every number traceable to a row in the data.
Clean, convert and join several files into one table, then compare groups and periods for rates, differences and trends.
Check a spreadsheet for duplicates, missing entries and bad dates before a client sees the file, with the row number of every problem.
Pull public data series from government sources such as the Census and the CDC, or get a ranked shortlist of published datasets from open-data catalogs.
Get the results back as charts, a spreadsheet and a written report that explains what the numbers mean.
Here are 18 months of orders. What is the story?
For this question I’ll go with find the stories in the data. Other ways I could go:
- Find the stories in the dataEvery finding keeps its calculationChosen
- Clean and combine the filesFix, convert, join
- Compare groups and periodsRates, differences, trends
- Build a model from assumptionsA scenario you can change
- Charts and a spreadsheetFiles people can use
- Explain what the numbers meanA written findings report
Some of the ways. Say the word and I’ll take another.
I’ll read the 18 months of orders, look for the stories, and keep the calculation behind each one.
Finding 1. Customers whose first order is a subscription come back within 30 days twice as often as customers who start with a single bag: 44% against 22%.
An order export, read and charted, with the calculation behind the finding.
Hand over your next job.
Your next campaign can start from a topic, a survey you want to run or a file you already have.