Research exhibit / Opportunities engine
Aurora
Which published opportunity is relevant enough for our team to investigate next?
A promising opportunity is useless if the source or deadline is wrong.
A list of notices can be difficult to qualify. A business needs to understand relevance, key dates, requirements and who should review the opportunity before choosing a next step.
Concept study · research not yet conducted. This is a specific investigation plan around an illustrative website. No interviews, verified search volumes, implemented software or measured outcomes are claimed.
The question shapes
the whole website.
This opportunity discovery direction would be useful if the visitor can answer the study’s question and the receiving team can act on the resulting information. That remains a hypothesis to test.
The proposed website responseConnect relevant notices to a clear opportunity workspace: show why an item was matched, its deadline, a review owner and an editable decision record. Keep the decision with the business; bid writing and submission are outside this idea.
Who is choosing?
What is at stake?
A useful profile describes the job, concern and evidence behind a decision. These are hypotheses to investigate, not invented personas presented as facts.
- Profile hypothesis · not verified
Business development teams
- The job
- Which published opportunity is relevant enough for our team to investigate next?
- The worry
- A list of notices can be difficult to qualify. A business needs to understand relevance, key dates, requirements and who should review the opportunity before choosing a next step.
- Proof to look for
- the original notice, source date, deadline, scope and the criteria used to judge relevance
Find the right participants Invite people who currently perform this role and have a recent relevant decision to discuss. Include different experience and access needs.
Verify or reject Ask for a specific recent example, trace it against permitted records and retain disagreements as findings. Record source, role, date and research limitations. - Profile hypothesis · not verified
Commercial reviewers
- The job
- Connect notice source, eligibility, dates, relevance notes and responsibility for a human qualification decision.
- The worry
- Will the information be sufficient to judge a relevant opportunity to review?
- Proof to look for
- the original notice, source date, deadline, scope and the criteria used to judge relevance
Find the right participants Invite people who currently perform this role and have a recent relevant decision to discuss. Include different experience and access needs.
Verify or reject Ask for a specific recent example, trace it against permitted records and retain disagreements as findings. Record source, role, date and research limitations. - Profile hypothesis · not verified
Opportunity coordinators
- The job
- An opportunities engine with explainable filters, a saved shortlist and a clear original-source link.
- The worry
- Can the receiving team act on the request without guessing missing facts?
- Proof to look for
- the original notice, source date, deadline, scope and the criteria used to judge relevance
Find the right participants Invite people who currently perform this role and have a recent relevant decision to discuss. Include different experience and access needs.
Verify or reject Ask for a specific recent example, trace it against permitted records and retain disagreements as findings. Record source, role, date and research limitations.
Start with people
Watch a team assess relevant published notices and ask which facts let them dismiss a poor fit or save a promising opportunity.
The proposed record: A short list of audience needs, recurring questions and assumptions that still need checking.
Search for participants
Start with permitted enquiries and customer records, service-owner referrals and relevant professional communities. Recruit against the role and task, include different experience levels and keep an appropriate consent record.
Keep an evidence trail
Record participant criteria, research date, exact task, supporting observations and contradictory findings. An assumption is accepted, revised or rejected against what people actually do and say.
Follow one need
all the way through.
Every useful part of the page should have a reason. Follow the connection from a buyer’s situation to the evidence, original words, search answer and test.
One decision, traced from beginning to end
Which published opportunity is relevant enough for our team to investigate next?
the original notice, source date, deadline, scope and the criteria used to judge relevance
Find a relevant notice. Check the original source.
find relevant business opportunities
A shortlist should make scope, source and deadlines easy to review. Explain why a notice may be relevant and direct the team to the original requirements. The team decides whether an opportunity merits further work.
Can the buyer recognise whether the service covers their need?
One decision, traced from beginning to end
The buyer needs evidence relevant to a relevant opportunity to review.
the original notice, source date, deadline, scope and the criteria used to judge relevance
The conditions and evidence behind a relevant opportunity to review
opportunity notice deadline
Before judging a relevant opportunity to review, check the original notice, source date, deadline, scope and the criteria used to judge relevance. Ask about the details that still need confirmation.
Can the buyer find the relevant constraints and supporting facts?
One decision, traced from beginning to end
The receiving team needs a usable handover.
Connect notice source, eligibility, dates, relevance notes and responsibility for a human qualification decision.
A useful brief for a relevant opportunity to review
opportunity shortlist source verification
Prepare the information needed for a relevant opportunity to review. The proposed receiving team would review the brief, identify missing facts and confirm the next step.
Can the buyer prepare the information the receiving team actually needs?
Follow the information
Connect notice source, eligibility, dates, relevance notes and responsibility for a human qualification decision.
The proposed record: One documented journey with required information, decision points and a named owner for each handover.
Find the words.
Check the intent.
Keywords are a research decision, not decoration. These case-specific candidates show why a term might belong, what page it needs and which evidence would confirm or reject it.
| Candidate and reasoning | Intent and page response | How to validate |
|---|---|---|
| find relevant business opportunitiesA candidate phrase for the decision about a relevant opportunity to review; investigate it with the people in this study, rather than choosing it for volume alone. | Understand the options Service overview and decision question | Compare buyer vocabulary, permitted search-query data, relevant search results and service-owner facts. Record contradictions. Search volume and ranking difficulty are not supplied.Relevance judgement |
| opportunity notice deadlineA candidate phrase for the decision about a relevant opportunity to review; investigate it with the people in this study, rather than choosing it for volume alone. | Check suitability Capability, evidence and constraints | Compare buyer vocabulary, permitted search-query data, relevant search results and service-owner facts. Record contradictions. Search volume and ranking difficulty are not supplied.Source verification |
| opportunity shortlist source verificationA candidate phrase for the decision about a relevant opportunity to review; investigate it with the people in this study, rather than choosing it for volume alone. | Prepare the next action Process explanation and enquiry requirements | Compare buyer vocabulary, permitted search-query data, relevant search results and service-owner facts. Record contradictions. Search volume and ranking difficulty are not supplied.Review effort |
Inspect the market
Review competing offers and alternative ways to solve the buyer’s problem. Record the public promises, qualifying details and unanswered questions. Investigate whether a gap matters to the buyer before making it the message.
Make the selection auditable
Keep the query source, intent, result-page review, target page, date and owner. Volume and competition remain unknown here. A real keyword tool estimate must include its tool, geography and period; first-party search data has its own coverage limits.
One question.
A proper, useful answer.
SEO helps people find a relevant page. AEO means structuring an answer people can understand and answer systems can interpret. Both begin with factual, specific content.
Editorial sketch · facts and wording still to verify
Which published opportunity is relevant enough for our team to investigate next?
A shortlist should make scope, source and deadlines easy to review. Explain why a notice may be relevant and direct the team to the original requirements. The team decides whether an opportunity merits further work.
Search foundations
A distinct title, relevant heading, readable text, useful internal links and a page that can be crawled and indexed when publicly launched.
Answer structure
Place the direct answer beside the question, then show conditions, evidence, responsible ownership and the appropriate action.
Measure each channel
Use permitted query, impression, visit and enquiry data. Review available AI-citation reports separately. A citation, a visit and an enquiry are different events.
Google describes optimisation for its generative features as SEO; it does not require a special AI file or AI-only schema. Rankings, inclusion in answers and commercial results cannot be guaranteed. This password-protected preview supplies no live search-performance evidence.
Write this business.
From the beginning.
Every published sentence should earn its place by answering a buyer’s question or explaining a verified fact. The content is developed from scratch around this brief; a generic AI paragraph cannot supply client knowledge or evidence.
Gather the real material
Observe bid teams making shortlist decisions and document the evidence behind each choice. Map the qualification checks, ownership and deadline information that an operator needs before interpreting any recommendation or ranking.
Find a relevant notice. Check the original source.
A shortlist should make scope, source and deadlines easy to review. Explain why a notice may be relevant and direct the team to the original requirements. The team decides whether an opportunity merits further work.
Approve the meaning
The owner reviews business facts and permissions; the research team checks the connection to the buyer question; likely users test whether the answer is usable. Unverified claims stay out.
- The buyer asks
- Which published opportunity is relevant enough for our team to investigate next?
- The evidence required
- the original notice, source date, deadline, scope and the criteria used to judge relevance
- Before it is published
- Trace each factual claim to an approved source, confirm permissions, check the meaning with the responsible owner and test comprehension with the intended audience.
The draft above is an illustrative content direction, not approved client copy. Original writing does not mean inventing facts. No claim is made that a buyer has approved these words.
Each part has a job.
Each job has a check.
This annotated blueprint shows the intended information hierarchy. It is a proposed content structure, not an eye-tracking heatmap or a measured analysis of the artwork.
Find a relevant notice. Check the original source.
Which published opportunity is relevant enough for our team to investigate next?
Relevant facts and evidence
the original notice, source date, deadline, scope and the criteria used to judge relevance
The process and its limits
Connect notice source, eligibility, dates, relevance notes and responsibility for a human qualification decision.
A useful request, with a clear next step
Opportunities engine
- A · Opening proposition
Which published opportunity is relevant enough for our team to investigate next?
The page decision Show the relevant problem, the service category and the intended buyer before decorative detail.
Evidence to obtain Observe bid teams making shortlist decisions and document the evidence behind each choice. Map the qualification checks, ownership and deadline information that an operator needs before interpreting any recommendation or ranking.
Check it Can the buyer recognise whether the service covers their need? - B · Capability and proof
What would help me judge a relevant opportunity to review?
The page decision Connect relevant notices to a clear opportunity workspace: show why an item was matched, its deadline, a review owner and an editable decision record. Keep the decision with the business; bid writing and submission are outside this idea.
Evidence to obtain Request and review the original notice, source date, deadline, scope and the criteria used to judge relevance.
Check it Can the buyer find the relevant constraints and supporting facts? - C · Process and boundaries
What happens, what is required and what remains uncertain?
The page decision Make the scope, dependencies and information needed for the next action visible in the page.
Evidence to obtain Connect notice source, eligibility, dates, relevance notes and responsibility for a human qualification decision.
Check it Ask the buyer to explain the process in their own words; note incorrect expectations. - D · Action and handover
What do I send and who takes it from here?
The page decision An opportunities engine with explainable filters, a saved shortlist and a clear original-source link.
Evidence to obtain Collect permitted published notices, match agreed criteria and explain why an item was flagged.
Check it Find a relevant notice for a fictional business, explain why it may fit and confirm its deadline on the original source. Check the same essential path with keyboard navigation and relevant assistive technology.
Build a small testable route
An opportunities engine with explainable filters, a saved shortlist and a clear original-source link.
The proposed record: A small public-facing route and a sketch of the associated staff task, ready to test with fictional records.
Visual hierarchy, recognition and scanning guide a design hypothesis. They do not read a buyer’s mind or prove a sales effect. Observe the actual task, including failures and access barriers.
The website starts it.
The working system carries it.
A well-designed enquiry can become useful information for the people doing the work. These are case-specific capabilities to scope, not a claim that the artwork hides a working backend.
- 01
Opportunities engine
Collect permitted published notices, match agreed criteria and explain why an item was flagged.
Permitted published notice explainable relevance check human qualification saved source and review date. - 02
Deadline and owner reminders
Give a saved opportunity a review owner and dates so the team can decide whether to pursue it.
Agreed date, record and owner reminder or exception staff review confirmed next action. - 03
Source and change log
Keep the publisher link and last-checked information visible, flagging material notice changes for human review.
Website request or agreed record named staff owner review and next action visible status for authorised people.
These are possible bespoke backend features, not a view into an existing working system. Scope, data, permissions, integrations and staff responsibilities would be agreed after discovery.
Opportunity discovery and qualification only. This concept does not offer bid writing, tender submission or automatic commitments.
Give a real task.
Watch what happens.
Set a believable goal, use neutral instructions and record success, confusion and failure. Observe likely buyers and receiving staff; include relevant accessibility checks.
Find a relevant notice for a fictional business, explain why it may fit and confirm its deadline on the original source. Check the same essential path with keyboard navigation and relevant assistive technology.
Observe a real task
Find a relevant notice for a fictional business, explain why it may fit and confirm its deadline on the original source. Check the same essential path with keyboard navigation and relevant assistive technology.
The proposed record: A record of task completion, misunderstandings and access barriers, with the scope and participants stated.
- Baseline to establish
Relevance judgement
Record whether the selected notice matches the stated business criteria.
Record a baseline first. Compare similar tasks after a change, include failures and note other factors that may explain differences. No result is claimed here. - Baseline to establish
Source verification
Check whether the user can verify current eligibility and dates with the publisher.
Record a baseline first. Compare similar tasks after a change, include failures and note other factors that may explain differences. No result is claimed here. - Baseline to establish
Review effort
Time the human qualification of comparable notices, recording rejected matches as well.
Record a baseline first. Compare similar tasks after a change, include failures and note other factors that may explain differences. No result is claimed here.
Record the conditions
Task wording, participant criteria, device, starting point, assistance, completion rule and failure reason all belong in the record. Agree them before observing the task.
Change one useful thing
Prioritise a clear barrier, make a specific revision and repeat a comparable task. Keep evidence that contradicts the hypothesis. Small usability samples reveal problems; they do not establish a population conversion rate.
The maths must match
the question.
A number becomes useful when its numerator, denominator, conditions and limitations are clear. These specifications connect the proposed measures to this case’s actual task.
Relevance judgement
Tasks meeting the agreed success criterion ÷ all attempted tasks × 100
- Unit
- Percentage of attempted tasks
- Collect
- State the success criterion first. Record successful, failed and assisted attempts; count every attempt in the denominator.
- Interpret
- Compare like-for-like tasks and report the counts as well as the percentage. A small sample does not establish a population conversion rate.
Source verification
Tasks meeting the agreed success criterion ÷ all attempted tasks × 100
- Unit
- Percentage of attempted tasks
- Collect
- State the success criterion first. Record successful, failed and assisted attempts; count every attempt in the denominator.
- Interpret
- Compare like-for-like tasks and report the counts as well as the percentage. A small sample does not establish a population conversion rate.
Review effort
Median of (finish time − start time) across comparable completed tasks
- Unit
- Minutes per completed task
- Collect
- Record the start event, finish event, pauses and whether the task was completed. Report unsuccessful tasks separately.
- Interpret
- A shorter duration is useful only if accuracy, completion and the receiving team’s work remain acceptable.
Show the calculation(8 ÷ 12) × 100 = 66.7%
No measured resultOne fictional arithmetic example: eight completed tasks out of twelve attempts. This is not a baseline, result, uplift, conversion forecast or measurement of this website.
| Measure | Completed | Attempts | Rate |
|---|---|---|---|
| Task completion | 8 | 12 | 66.7% |
This arithmetic demonstrates a method; it is not an improvement, forecast, estimate of revenue or proof of causation. A commercial claim needs suitable real data, an appropriate comparison and enough observations for its purpose.
Improve in small increments
Agree a starting measure for relevance judgement, source verification, review effort. Test the riskiest step first, change one clear problem, then repeat a comparable task; record other changes that could affect the comparison.
The proposed record: An agreed comparison, the next useful change and a note of what remains uncertain.
A proposed evaluation plan for an illustrative website study. No interviews, analytics results, user tests or customer outcomes are claimed for this concept.
The explanation is only useful
if the page works.
Set performance budgets, test relevant devices and check keyboard, zoom, labels and assistive technology. Build quality supports the task; a score alone does not show that a buyer can do it.
- Loading · LCP≤ 2.5 s
How quickly the main content appears.
- Responsiveness · INP≤ 200 ms
How promptly a page responds to interaction.
- Visual stability · CLS≤ 0.1
How much the layout unexpectedly shifts.
Published Core Web Vitals “good” thresholds at the 75th percentile of real visits, assessed separately for mobile and desktop. Proposed field targets, not measured scores for this concept. A lab test alone is not field evidence; user testing alone does not establish accessibility conformance.
What would persuade us
to keep or change it?
Research is useful when it changes a decision. This ledger connects the claims made in the brief to evidence that would support, refine or reject them.
| Claim to investigate | Evidence to gather | Decision rule |
|---|---|---|
| Buyer profile | Watch a team assess relevant published notices and ask which facts let them dismiss a poor fit or save a promising opportunity. | Use the profile only where the recorded examples support it; retain exceptions and unresolved assumptions. |
| Chosen phrase and intent | Compare the candidate phrases with buyer wording, permitted queries and relevant search results. | Record intent, source, date, suitable page and a reason to retain or reject each phrase. |
| Published content | Trace service statements to the original notice, source date, deadline, scope and the criteria used to judge relevance. | Write the answer from scratch, record factual approvals and remove unsupported claims. |
| Proposed improvement | Find a relevant notice for a fictional business, explain why it may fit and confirm its deadline on the original source. Check the same essential path with keyboard navigation and relevant assistive technology. | Establish a baseline, count failed attempts and compare agreed tasks before claiming any improvement. |
Published B2B research informs the questions. It does not prove this design will increase enquiries or sales.
Read the 16 sources, their scope and their limits
- 2025 B2B Buyer Experience Report (source, new tab)
Commercial B2B buyer survey
94% of surveyed buyers said they ranked their shortlist before speaking with sellers.
Nearly 4,000 buyers in North America, APAC and EMEA; qualifying purchases were at least $25,000 and the median was $200,000–$300,000.
Retrospective reports about higher-value B2B purchases; not a website experiment or a result for these fictional businesses.
- B2B Website Usability: public research overview (source, new tab)
Publisher’s public overview of its third-edition B2B research
Business websites need to support comparison, shared purchasing decisions and customer tasks after a purchase.
The public overview describes representative business users testing 293 B2B sites across the US, UK and Singapore, using several research methods.
The full paid report was not accessed. This overview does not supply a conversion estimate or validate any concept shown here.
- Learning about users and their needs (source, new tab)
Public-service research guidance
Investigate what people need to do, their current approach and the problems they experience; treat untested internal opinions as assumptions.
Guidance for government services, including people who use, provide or support them.
A transferable practice method, not a B2B commercial-outcome study.
- Creating an experience map (source, new tab)
Research synthesis and journey-mapping guidance
Combine observed user experiences to reveal stages, dependencies, pain points and handovers across a service.
Particularly relevant to journeys involving several events, people, teams or touchpoints.
The map must be grounded in research. A proposed diagram is not an observed customer journey.
- Making prototypes (source, new tab)
Design practice guidance
Explore and test alternative designs with a small prototype before committing to a full build.
Includes sketches and realistic interactive prototypes for desktop and mobile tasks.
A convincing prototype is not proof of production security, reliability, integration or business value.
- Using moderated usability testing (source, new tab)
Task-based usability research guidance
Observe likely users attempting believable tasks to uncover confusing language, layout and interaction.
Tasks should have a clear goal without revealing the answer; sessions require appropriate recruitment and a consistent discussion guide.
Small qualitative rounds reveal specific problems. They do not justify general population percentages or guaranteed conversion gains.
- Using performance data to improve your service (source, new tab)
Performance measurement guidance
Plan measurement from the beginning and combine task data with research to decide what to improve.
The government service metrics are accompanied by guidance to choose additional measures appropriate to each service.
Commercial projects need their own measures. A before-and-after change may have several causes and is not automatically causal evidence.
- Involving Users in Evaluating Web Accessibility (source, new tab)
Accessibility evaluation guidance
Testing with disabled users can reveal practical barriers; combine this with standards-based evaluation and report the review’s limits.
Covers early consultations, formal studies, assistive technology and cautious interpretation of limited samples.
One user’s experience does not represent every disability. User testing alone cannot establish WCAG conformance.
- Search Engine Optimization (SEO) Starter Guide (source, new tab)
Primary-source practice guidance
Create original, readable and well organised content; anticipate the different terms experienced and new readers may use; give related pages descriptive links and relevant images useful context.
Official introductory guidance for search discovery and presentation. Search systems can understand related language without every exact phrase being repeated.
No guaranteed ranking, indexing or snippet wording. Avoid keyword stuffing, arbitrary word counts and invented search volumes. Public discovery and search eligibility need their own checks; publication does not guarantee a ranking or AI citation.
- Influencing your title links in search results (source, new tab)
Primary-source practice guidance
Give each page a concise, descriptive and distinct title, avoiding repeated boilerplate and keyword stuffing.
Official guidance explains how Google selects search-result titles using several on-page and off-page sources.
A written title is a preference signal, not control over the exact displayed search title or a promised click-through rate.
- Creating helpful, reliable, people-first content (source, new tab)
Primary-source practice guidance
Google encourages original, useful, reliable content for an intended audience. Its current quality guidance discusses effort, originality, skill and factual accuracy, with clear attribution where appropriate.
Publisher self-assessment guidance. Google says SEO is useful when applied to people-first content rather than search-engine-first content.
This is not a universal content checklist that guarantees ranking or sales. E-E-A-T is not itself a specific ranking factor. Each word cannot be separately scientifically validated.
- Optimizing your website for generative AI features on Google Search (source, new tab)
Primary-source practice guidance
Google treats AEO/GEO work for its generative search experiences as SEO. It prioritises original, useful content and ordinary technical structure; special AI files, schema and content chunking are unnecessary.
Google Search. Eligibility includes indexing, snippet eligibility and current Search Console generative-feature inclusion settings.
No guaranteed indexing, citations, position, sales outcome or eligibility in other services.
- Introducing AI Performance in Bing Webmaster Tools Public Preview (source, new tab)
Primary-source practice guidance
Bing recommends clearer structure, evidence, expertise, accurate current information and consistent meaning across text and media. Its preview reporting shows citation activity across supported Microsoft AI experiences.
Bing, Microsoft Copilot and selected supported integrations. Citation counts and sampled grounding queries describe supported answer experiences.
The cited-page measures do not indicate ranking, authority, placement or sales. Readable answers and references do not guarantee inclusion in an AI response.
- How People Read Online: New and Old Findings (source, new tab)
Publisher’s synthesis of eyetracking research
People commonly scan online content while looking for relevant information. Reading patterns vary with layout, task, motivation and the individual.
Public summary of 13 years of eyetracking research, over 500 participants and more than 750 hours of observation.
Not every visitor scans all content, and no single layout pattern fits all pages. The public article does not establish a revenue increase for this project; the paid report was not accessed.
- Memory Recognition and Recall in User Interfaces (source, new tab)
Cognitive psychology and interface-design guidance
Visible choices and contextual cues can support recognition instead of asking people to recall information unaided.
A cognitive-psychology explanation applied to menus, search and interface content; published January 2024.
A design heuristic, not proof of universal behaviour or a specific conversion increase. Labels still need testing with the intended audience.
- Web Vitals (source, new tab)
Published web-performance measurement guidance
The published good-experience thresholds are LCP at most 2.5 seconds, INP at most 200 milliseconds and CLS at most 0.1.
Assess the 75th percentile of real page loads, separately for mobile and desktop, across all three Core Web Vitals.
These are performance targets, not measured results for the concepts. A laboratory result is diagnostic and is not equivalent to representative field data or a sales forecast.