Content strategy · 100 content ideas

What does 100 content ideas mean in practice?

A useful set of 100 content ideas is a governed inventory of audience questions, evidence, formats, and intended outcomes—not a promise to publish 100 pages. Build candidates from defined audience jobs, attach a source and decision to each, merge duplicate intent, score the remainder, and advance only the ideas your team can answer credibly.

Treat 100 as an exploration target, not a publishing quota

The phrase “100 content ideas” is most useful as a prompt to explore a problem space broadly. It should produce a candidate inventory from which a much smaller number of briefs may be selected. A candidate is not automatically a keyword target, URL, article, campaign, or commitment. Give every row a stable ID, an audience, a specific question or job, a proposed evidence source, an intended outcome, an owner, and a disposition such as explore, merge, defer, reject, test, or brief. This distinction prevents brainstorming volume from becoming scaled publication by default. It also makes the list auditable: another editor can see why two similar phrases belong to one answer and why an appealing topic was rejected when the organization lacked evidence. Reaching exactly 100 has no inherent business or search value; the useful result is a sufficiently varied, source-aware inventory that improves decisions.

Start with an objective and a defined audience decision

Write one sentence describing what the content program should help a specific audience understand, decide, or do, and how the team will know whether that happened. Then note constraints such as time, budget, expertise, review risk, accessibility, and available channels. The UK Government Functional Standard for communication says objectives should clarify what is required and align campaign components; it also describes audience insight as understanding attitudes, habits, and preferences so communication can be relevant and meaningful. Use that official framework as planning evidence, not as a claim that its rules apply to every private organization. For this exercise, an objective such as “help first-time content leads select a defensible research method” is more useful than “get more traffic.” It narrows who the inventory serves and gives reviewers a reason to remove ideas that cannot contribute to the decision.

Create five evidence-based audience slices

Choose five audience slices that matter to the same program, using first-party research where available. A B2B product team might use new evaluators, active practitioners, operational owners, economic buyers, and existing customers. These are working hypotheses until supported by interviews, support themes, product research, sales notes, or other appropriate evidence. Avoid inventing personas from job titles alone or treating public posters as representative customers. For each slice, capture the situation, knowledge level, constraint, decision authority, and vocabulary it actually uses. Record contrary evidence too: if two roles share the same task and need the same answer, keep one content intent rather than manufacturing separate pages. Five slices provide one dimension of the 100-candidate matrix, but teams may use fewer when their evidence does not support five meaningful differences. The point is coverage of distinct audience contexts, not filling a spreadsheet with cosmetic labels.

Map four jobs or uncertainties for each audience

For every audience slice, write four concrete jobs or uncertainties. Useful categories include understanding a concept, comparing approaches, completing a task, and evaluating a result. Phrase each as a question a qualified person could reasonably ask, then note the decision that a satisfactory answer would enable. Twenty audience-job pairs now exist: five slices multiplied by four jobs. Check each pair against research notes and the existing content corpus. Merge synonyms, remove questions already answered well, and reject questions outside the organization's expertise. Keyword data may help order investigation, but it does not prove that the question represents customer demand or deserves a page. A visible public question can be a useful lead, yet it remains an observation from bounded evidence. This step deliberately converts a vague topic list into explicit intent, which makes later duplication and evidence checks possible.

Generate five answer treatments without creating five URLs

Apply five possible treatments to each of the twenty audience-job pairs: a direct explanation, a worked example, a checklist, a comparison, and a diagnostic or measurement note. The arithmetic yields 100 candidate treatments. It does not mean all combinations are good or that each requires a standalone page. Several treatments may become sections of one canonical guide; a checklist might become a downloadable or on-page template; a diagnostic might belong in product documentation; and unsupported combinations should be rejected. Include a proposed proof type for every treatment, such as first-party procedure, product documentation, an official standard, a named subject-matter review, or original analysis with disclosed method. If no suitable evidence can be identified, label the candidate “evidence missing” rather than drafting confident prose. This matrix creates variety while retaining a deterministic relationship between audience, job, treatment, and proof.

Use source-linked conversations as hypotheses, not market totals

Public conversations can contribute wording, objections, edge cases, and candidate questions. Preserve the original source URL, date, matched term, available public metric, and the interpretation you propose. What's Trending's first-party API guide documents ranked topics and source-linked public evidence, explains that metric fields vary by source, and says missing values mean unavailable rather than zero. It also instructs users to inspect original URLs before turning a trend into a claim. Those boundaries should follow each candidate into the inventory. Do not convert a handful of posts into a prevalence percentage, a universal customer preference, or a revenue forecast. Where the consequence matters, validate the hypothesis with suitable first-party audience research, product data, or expert review. A source trail improves inspectability; it does not make the observed sample representative or prove the proposed content will perform.

Score evidence, usefulness, distinctness, risk, and effort

Score each candidate with explicit, inspectable rules. One practical rubric assigns zero to two points for audience evidence, organizational expertise, decision usefulness, distinctness from existing answers, and measurement feasibility, then subtracts zero to two points each for factual risk and maintenance cost. Keep the component scores beside the total so a reviewer can challenge the judgment. Treat hard gates separately: a candidate fails regardless of score if it duplicates canonical intent, requires an unsafe or unprovable claim, exposes private information, lacks an accountable owner, or cannot identify a credible evidence path. Sorting helps allocate research; it does not predict ranking, reach, conversion, or revenue. Review clusters as well as rows, because ten individually plausible candidates can still form a repetitive doorway set. Record merge and rejection reasons so the same weak variants do not reappear in the next brainstorming session.

Turn selected candidates into people-first briefs

Advance only the highest-supported candidates into briefs. Each brief should contain the direct question, intended audience, useful answer boundary, evidence ledger, material claims, practical example, limitations, internal-link context, measurement plan, and the canonical intent it owns. Google Search Central recommends content created primarily for people, with original information or analysis, clear sourcing and expertise, and a satisfying answer for an intended audience. It warns against producing many topics mainly to attract search visits and against relying on extensive automation without added value. Apply that guidance as an editorial quality check, not a ranking formula. One complete guide that resolves a task may responsibly absorb several matrix treatments. Conversely, leave a valuable question unassigned until qualified evidence is available. A large idea bank is successful when it improves selection and answer quality, even if only a few items become publishable content.

Evaluate outcomes separately from publishing activity

Define how each selected brief will be evaluated before production. AMEC's Integrated Evaluation Framework distinguishes outputs, audience out-takes, outcomes, and organizational impact. In practice, a published page and its distribution are outputs; observed attention or understanding may be out-takes; a verified audience action may be an outcome; and contribution to an organizational objective is a separate impact question. Do not label the number of ideas, pages, impressions, or engagements as business impact. Choose a primary measure appropriate to the content's job, retain the underlying counts and time window, and document attribution and data gaps. Compare results with the original objective and use the evidence to update the inventory: expand a well-supported question, merge overlapping answers, revise an unclear treatment, or stop maintaining a low-value asset. The framework organizes learning but does not by itself establish causation or guarantee performance.

Maintain the inventory as a decision log

Review the 100 candidates on a fixed cadence rather than generating a fresh list whenever the calendar looks empty. Update evidence dates, owners, dependencies, risk, and disposition. Re-run exact intent checks against live and draft content before a candidate advances. Preserve earlier decisions instead of silently deleting them, and assign a reason when an item changes state. New audience evidence may split one question; stronger canonical content may merge several; a product change may invalidate a proposed answer. Avoid changing dates or reviving old ideas merely to appear current. The inventory should expose uncertainty and maintenance burden, not hide them. A compact backlog of defensible, distinct questions is preferable to 100 nominally “ready” titles. The number is a repeatable exploration device; publication remains a separate process with factual, privacy, quality, and operational gates.

Examples

A small analytics company wants ideas for operations leaders evaluating a new reporting workflow. It defines five evidence-backed audience slices, including first-time evaluators and implementation owners, and four jobs for each: understand, compare, implement, and evaluate. The team applies five treatments to every pair, creating 100 rows. One row proposes a separate article for every wording variant of “weekly reporting checklist.” Exact intent review merges those variants into a single canonical checklist. Another proposes a benchmark the company cannot substantiate, so it is marked evidence missing. A public discussion supplies wording for a permissions concern, but the source is retained as a bounded observation and the team validates the workflow with product documentation before drafting. Twelve candidates pass the hard gates, four become sections in one guide, two become templates, three remain research questions, and three are deferred. The team measures qualified task completion separately from page output and preserves all dispositions for the next review. The exercise produced 100 inspectable candidates, not 100 publishing commitments.

Limits and interpretation

The 5 × 4 × 5 matrix is a recommended ideation method, not an official standard, and teams should use fewer dimensions when their evidence cannot support meaningful distinctions.
A list of 100 candidates does not demonstrate 100 separate search intents, customer needs, or publishable URLs; exact-intent deduplication and canonical consolidation remain required.
Keyword volumes, public posts, impressions, and engagement are bounded research or diagnostic inputs. They do not by themselves prove representative demand, causation, revenue, or business impact.
Official planning and evaluation frameworks organize decisions, but they do not supply an organization's audience evidence, expertise, attribution method, or performance guarantee.
What's Trending documents selected public-source evidence with varying metric availability; it does not claim exhaustive coverage, private-data access, representative sampling, or guaranteed content results.

Should a team publish all 100 content ideas?

No. Treat the 100 rows as candidates. Merge duplicate intent, reject unsupported or unsafe claims, cluster related treatments into stronger canonical answers, and brief only the items with an audience need, credible evidence, accountable owner, and useful decision path.

How can 100 content ideas stay distinct?

Give each candidate an explicit audience, question, decision, treatment, and proof type, then compare normalized intent with the existing corpus. Preserve modifiers that change the job, but merge cosmetic wording and record the owning canonical plus the audit reason.

Can social listening generate 100 validated content ideas?

It can surface source-linked language and hypotheses for an inventory, but it cannot validate market-wide demand or business value by itself. Inspect original sources and combine bounded observations with appropriate first-party research, product facts, and outcome measurement.