How to Build a Source-Backed PhD Search System

Most PhD searches do not fail because the applicant cannot find enough links. They fail because the links never become a usable record. A vacancy is saved without its closing date. A scholarship is copied from a search result but not connected to the university that administers it. A promising research group is mixed into the same list as an open, funded position. A note from last year sits beside a current call, and after a few weeks it is hard to tell which information still applies.

This is not just a filing problem. It changes the decisions you make. If you cannot tell which page controls a claim, you may prepare for a route that is already closed, miss a separate nomination step, or spend time on a project that does not match your background. A long list can create the feeling of progress while leaving the important questions unanswered.

A source-backed search system gives every lead a place, an owner, a source, a date, a fit note, and a next action. It separates discovery from verification and verification from the final decision. It also keeps uncertainty visible. When a fact is not stated, the record says “not stated” instead of quietly turning a guess into a fact.

This guide is for applicants looking for STEM PhD opportunities abroad, including project vacancies, doctoral programmes, training partnerships, networks, scholarships, and individual fellowships. It is also useful if you already have a spreadsheet or a folder of bookmarks but cannot trust it. The guide does not verify an individual vacancy, assess your personal eligibility, predict admission, or promise funding. For those decisions, use the current page owned by the university, employer, funder, or government authority.

The short answer is to build one small, repeatable system rather than one huge list. Define a search scope, capture a lead without losing its original context, move it through discovery and verification stages, keep evidence beside each important claim, review freshness, and record why a lead became active, paused, or closed. The system should make the next decision easier to see. It should not become a second full-time job.

A search system is more than a saved-link folder

A saved link only tells you that a page caught your attention. It does not tell you what the page is, who owns it, when you checked it, or what you intend to do. Those missing pieces are the reason search lists become unreliable.

Consider three records that might all be called “PhD opportunity” in a browser folder:

• an open project vacancy published by a university employer;
• a doctoral training partnership that is relevant but has no current applicant call; and
• a research group whose recent publications fit your interests, with recruitment status unknown.

All three may deserve a place in your search. They do not deserve the same status. The first may need an application decision today. The second may need monitoring for a provider-led call. The third may need research reading or a carefully timed enquiry. If you put them in one undifferentiated list, urgency and uncertainty become confused.

Your system should answer five questions for every lead:

The record is useful even when the answer is “not open” or “not stated”. A closed call with a clear source can become a monitoring item. A page with unclear ownership can be removed before it costs you time. A research group with no vacancy can remain a research lead without being presented to yourself as a funded position.

The four layers of a reliable record

Keep four kinds of information separate inside each record.

The first layer is the source fact. This is what the current page says about the route, project, host, funding, eligibility, application process, or date. Use the source's language where precision matters, and link to the exact page rather than to a search page.

The second layer is the source context. Record who published the page, what kind of page it is, when it was issued or updated if that information is shown, and whether it links to a more specific instruction. A university vacancy, a funder call, and a department news item can all mention the same project but have different authority for different claims.

The third layer is your fit. Write your own assessment of the research question, methods, academic preparation, language evidence, location, timing, and practical constraints. Mark it as your note, not as a statement made by the institution. This distinction prevents a personal impression from becoming an apparent requirement.

The fourth layer is the open question and next action. A lead can be relevant while still needing a direct answer. Record the person or office that can answer it, the question in plain language, and what you will do if the answer is not available.

When these layers are mixed, a sentence such as “international applicants accepted and fully funded” may combine a general university statement, a project phrase, and your own assumption. When the layers are separate, you can see exactly which part needs checking.

A source-backed principle

The source-backed principle is simple: every decision-worthy claim needs a source that can control that decision. A discovery service can help you locate a vacancy. It cannot rewrite the employer's conditions. A funder overview can explain a scheme. It cannot decide whether your university will nominate you. A researcher's profile can show a publication or affiliation. It cannot prove that a doctoral place is available.

The DAAD guide to finding a PhD position in Germany is a useful example of a discovery guide that directs applicants towards institutions and opportunities. EURAXESS Jobs is another discovery route for research vacancies. Use both to find names and routes, then move to the current page of the university, employer, institute, or funder that controls the next step.

This rule also helps with search-engine results. A result snippet may contain a useful phrase but can be shortened or drawn from a paragraph that has changed context. Open the page and check its title, owner, route, and linked instructions. If the page is not clear, treat it as a lead to investigate, not as evidence to copy.

Set a focused search scope before you collect opportunities

The first practical step is not opening another search tab. It is deciding what you are willing to search for. A focused scope reduces noise and makes comparison possible. It also protects you from changing your criteria every time a new result looks attractive.

Your scope does not have to be permanent. It is a working boundary for one search cycle. Write it in a short paragraph and give it a date. You can revise it when your evidence changes, but you should be able to explain why it changed.

Start with the research problem, not only the degree name

“PhD in computer science” is too wide to guide a serious search. Add the question, problem, system, method, or kind of evidence that matters to you. The description can be provisional. It is a search instrument, not a final proposal.

For example, a broad interest in energy systems can become a search note about fault detection in renewable-energy infrastructure using machine-learning methods and field data. A broad interest in biology can become a note about microbial responses to pollutants using experimental and sequencing methods. The point is not to make the wording impressive. The point is to give yourself terms that appear on research-group, project, and funder pages.

Write three versions of the same interest:

Use the three versions together when searching. If you use only a degree title, large generic directories will dominate. If you use only a method, you may miss the programme or application language used by the host. The combination keeps discovery connected to a real research context.

Set a degree and career boundary

State whether you are looking for a PhD, an integrated doctorate, a doctoral researcher employment route, or another doctoral-level path. Do not mix postdoctoral positions into the same active queue. A postdoc can be a useful future route, but it has different eligibility, evidence, and application expectations. Keep it in a separate monitor list if it helps you plan.

Define whether you need a structured doctoral programme, a defined project, a network, or any of these routes. You do not need to choose only one category at the beginning. You do need to label the category so that you do not compare a project vacancy with a broad programme as if they were identical offers.

If you are still deciding between routes, record the decision you need to make. For example: “I will consider both project-led and programme-led PhDs, but I need a source that explains how admission and funding connect.” That sentence gives your search a question instead of an unbounded promise to look at everything.

Use geography only when it changes the decision

Country is often used as a search keyword even when it does not answer a research or application question. Include a country or region when it affects a route you need to understand, such as a government funding scheme, a language requirement, a mobility condition, or a realistic start window. Do not create a country list just to make the search feel complete.

A location can affect fee status, immigration evidence, employment conditions, travel, language, supervision, and access to data or facilities. These are real reasons to include it. “This country is popular” is not a search criterion.

If your geography is open, state that clearly and use a small set of filters such as research fit, funding evidence, and application route. An open geography can be a good choice when your method or field is specialised. It becomes unmanageable only when every new link is treated as equally relevant.

Set a timing window without inventing a deadline

Write the earliest start period you could accept, the latest you could start, and any hard personal constraints. Then search for routes that can plausibly fit that window. Do not copy a date from an old call or a discovery page and treat it as a current deadline.

Timing has several parts: the date you can submit, the date the university expects to enrol you, the funding period, any nomination stage, and the time you need for documents. A lead can appear open while its intended start is not workable for you. Record both the source timing and your own constraint.

Use “current date stated” and “date not stated” as different records. If a page does not state a closing date, do not fill one in from the date on a search result. Follow the application instructions or contact the named office if the timing affects your decision.

Write a scope statement

Put the scope at the top of your system. A useful version might read:

I am looking for STEM PhD routes connected to computational methods and applied energy research. I will consider project vacancies, structured programmes, and doctoral networks. I need a current source for the host, application route, funding terms, and international eligibility. I can begin within the next two academic intakes and will keep uncertain research leads separate from active opportunities.

Your statement will be different. What matters is that it tells you what belongs in the active search, what belongs in monitoring, and what should be left out.

Capture an opportunity without losing its original context

Once the scope is clear, build the record at the moment you discover a lead. Do not rely on opening ten tabs and promising yourself that you will document them later. The first capture can take two minutes. Reconstructing the same context after the page changes can take much longer.

Record the identity exactly

Copy the title as it appears on the controlling page. Add the institution, department or unit, project or programme name, and any reference number. If the page gives a host, employer, or funder, record each separately. A shortened title such as “AI PhD” is not enough to find the same route again.

Save the canonical application or opportunity URL. If you reached the page through a directory, save that discovery URL in a separate field, but do not use it as the main link. Keep the page that receives the application or states the current terms at the centre of the record.

Record the source type. Examples include current vacancy, programme page, funder call, provider scholarship page, network project page, university research page, or researcher profile. Source type gives you a quick sense of what the page can and cannot prove.

Capture the five core claims

At the first pass, record five claims only. This keeps the capture light while protecting the decisions you will later need to make.

The first claim is what is being offered: a project, programme place, studentship, scholarship, fellowship, employment role, or research lead. The second is who owns the next decision. The third is the application route and any separate step. The fourth is what the page says about funding or support. The fifth is the stated timing.

Write each claim as a short sentence and add the source link beside it. If the page does not answer one of the five, write “not stated”. This is faster and safer than drafting a summary that sounds complete.

For a vacancy, a first capture might say that a university department is recruiting for a named project through its employment portal; the page describes the research and says support is available, but it does not yet state the fee category. That record is enough to send the lead to verification. It is not enough to call the opportunity fully funded.

Add a fit note in your own words

After the source claims, write a separate fit note. Mention the research question, methods, data, preparation, and constraints that matter to you. Use “I” or “my notes” so the reader of the record can see that this is your judgement.

Avoid writing “the university wants…” unless the page actually says it. Use “the project page describes…” for source facts and “my background includes…” for personal evidence. The difference sounds small, but it stops your interpretation from becoming a false requirement.

Record both strong fit and possible mismatch. A useful note might say that your methods match the stated analysis approach, while your previous degree has a different title and needs an equivalency check. This is more useful than a single green tick.

Keep the original page and a short excerpt

Links can move. Save the page title and a short, decision-relevant excerpt or paraphrase with the URL. Do not copy an entire page into your notes. The goal is to preserve the claim and help you find the current source again, not to create a duplicate that will become stale.

If the page is a PDF, record the document title and the page number for a key statement. If it links to terms or a separate application form, save those links as related evidence. A vacancy page and its attached terms may together explain the route; one without the other can be misleading.

Assign an initial state

Use a small set of states that reflect decisions, not moods:

Do not use “saved” as the only state. It describes storage, not progress.

Use discovery, verification, and decision stages

A three-stage system keeps the search moving without making every lead urgent. The stages are connected, but they have different standards.

Discovery: collect names, not promises

Discovery is where you search broadly enough to find hosts, projects, programmes, funders, groups, and routes. It is fine for a discovery record to be incomplete. It is not fine to present it to yourself as a verified opportunity.

Use research questions, methods, group names, programme terms, funder terms, and official discovery services. The EURAXESS Jobs search can help locate European research vacancies. A DAAD discovery guide can help applicants understand ways to find positions. The MSCA Doctoral Networks page can help you understand a network route before you look for an individual project vacancy.

At discovery, ask only whether the lead deserves a record. Capture identity, source type, and one reason it fits your scope. Do not spend an hour analysing every document before you know whether the route is real or current.

Verification: test the claims against controlling sources

Verification begins when a lead has a plausible owner and a direct page. Read the current vacancy, programme, award, and relevant admissions pages. For each decision, ask which organisation controls the answer.

Check at least these areas: what the opportunity is, who can apply, what evidence is required, what support is stated, what the application route is, what date or stage applies, and which questions remain open. When the lead is international, keep immigration in a separate evidence track and use the current government source when it becomes relevant.

The UKRI studentships and doctoral training guidance illustrates why a funder page may point applicants to a provider. The UKRI doctoral training framework describes an investment structure, but a provider's current recruitment page controls the route for a student. Keep both in the record with their different roles.

For an individual fellowship, the current funder call controls the award conditions. MSCA Postdoctoral Fellowships is a postdoctoral example, not a doctoral checklist, but it shows why the call documents matter for a named action. Use the same habit for doctoral awards: read the actual call, not only the programme overview.

Decision: choose one next action

The decision stage does not mean that you must apply. It means that you decide what to do with the evidence you now have. Possible actions include apply, ask a precise question, wait for a named call, move the lead to monitoring, pause while obtaining a document, or close it.

Give the action a reason. “Apply because the source-backed route fits my scope and the required evidence is available” is better than “apply because it looks good.” “Pause because the current page does not state the international fee treatment” tells you what must change before you spend more time.

If you decide to apply, move the record to your application workflow rather than trying to manage every document in the search table. Link the application record to the opportunity record so that the source facts remain visible while your drafts and submissions develop.

Define hand-offs between stages

The system works when each stage has an exit rule. A discovered lead moves to verification when the direct source, host, and route are identified. A verified lead moves to active candidate when the core claims are supported and the fit is strong enough for a decision. An active candidate moves to the application workflow when you have chosen to submit and know the required steps.

If a lead cannot meet an exit rule, do not force it forward. Keep it in the earlier stage with one open question. This prevents the common pattern in which a lead is labelled “active” simply because it was interesting.

Build the search record around decisions

The record can be a spreadsheet, a database, a notes system, or a product workspace. The tool matters less than the fields and the habits. Start with enough structure to prevent errors, not enough structure to create data-entry work.

The identity block

Keep the identity fields at the top:

These fields protect against duplicate records. Two pages may describe one route. If you record the reference number and host, you can merge them rather than creating two apparent opportunities.

The route and status block

Record who recruits, who admits, and who administers support. They may be the same organisation, but never assume that they are. Add the application portal, whether a separate programme application is required, and whether a nomination or supervisor confirmation appears in the instructions.

Add the current state, the date you last checked the source, and the next review date you choose. A yearly review can be enough for stable research context. An open call, funding condition, eligibility rule, or immigration-related question needs closer attention, often before you act. The exact cadence should follow the risk and the source, not a decorative freshness label.

The eligibility block

Record the source wording for degree level, subject background, language evidence, residency or nationality condition, mobility rule, career stage, and any required documents. Do not reduce a complex condition to a green tick before you have checked the evidence.

Use three statuses for each item: confirmed from the current source, requires direct confirmation, or not stated. “Not stated” is not the same as “not required”. It simply tells you that the current page did not answer the question.

If a provider or funder uses an eligibility guide, link it beside the vacancy or programme page. A discovery listing may omit a condition that appears in the call. Keep the sources connected.

The funding block

Separate funding categories so that the word “funded” does not hide gaps. Record what the source says about tuition or fees, living support, employment or stipend form, period, research costs, travel, insurance, and other charges. If a category is not mentioned, mark it as not stated.

The NSF Graduate Research Fellowship Program is a useful example of why applicants must read the current programme page and solicitation for an individual award. The programme overview helps you find the route. The solicitation controls the detailed conditions. For a university studentship, the provider's award terms and fee information play the same role.

Do not add a number from a third-party summary to fill a blank. If the amount or coverage matters, follow the official source or ask the named provider.

The fit block

Record your fit in several short notes rather than one score. Include the research problem, methods, evidence or data, preparation, supervisor or group context, training needs, and practical constraints. Write what you know and what you need to learn.

Fit is not a ranking. A lead can be a strong scientific match and still be unworkable because the route is closed or the funding does not cover the relevant fee category. Conversely, a route with a broad project can be worth keeping while you clarify a method or supervisor question.

The evidence and open-questions block

For each important claim, store the source, page title, check date, and a short note. Then keep open questions in a separate list with the office or person who can answer them. A question such as “Does this award cover the international fee category for this route?” is actionable. “Is it really funded?” is too broad.

Add a next action to every open question. If the question is not important enough to resolve, remove it from the active record instead of letting it remain as permanent noise.

The application hand-off block

When you move a lead to an application workflow, record the submission route, required documents, reference tasks, document progress, and the next deadline exactly as stated by the current portal or instructions. Keep application-specific versions in the application record, not in the general search table. For the work after this hand-off, continue with the AcademicWings application guides, which cover preparation and submission rather than the search record itself.

The search record should remain a stable description of the opportunity. Your statement, proposal, CV, and referee messages will change as you prepare. Mixing them into one record makes it hard to know which version was submitted.

A product-neutral search-record template

Use the following template for every lead. You can copy it into a spreadsheet or notes page and adapt the fields to your situation.

Source and identity

Record name:

Opportunity type:

Host and department:

Funder, provider, or employer:

Reference number:

Current controlling URL:

Discovery URL:

Page title and document name:

Date found:

Last date checked:

Route

What is being offered?

Who recruits or receives the application?

Who decides academic admission?

Who administers or provides support?

Is a separate programme, funding, nomination, or supervisor step required?

Application portal and current stage:

Evidence

Academic and language requirements stated by the source:

International or route-specific conditions stated by the source:

Funding categories stated by the source:

Dates or stages stated by the source:

Documents and references stated by the source:

Unknown or conflicting points:

Personal fit

Why this research fits my question or method:

What preparation I can show:

What I need to learn or confirm:

Constraints that affect my decision:

Decision

Current state:

Next action:

Owner of the next action:

Date to revisit:

Reason for apply, monitor, pause, or close:

This template is intentionally plain. It asks for evidence before decoration. A colour or score can be useful later, but it should summarise a record rather than replace it.

Manage sources and freshness without creating fake certainty

Source management is not only about collecting more links. It is about knowing whether the link still answers the question you are asking.

Use a source hierarchy

Start with the source closest to the decision. For a live vacancy, this is often the current university, employer, or institute page. For a doctoral training award, it may be the provider's current call and application instructions, connected to the funder's framework. For an individual fellowship, it is the current funder call and host instructions. For immigration, it is the official government authority.

Use discovery services, journal directories, researcher profiles, and search engines to locate the source. The Directory of Open Access Journals can help you discover journals or field information, but inclusion is not a complete quality judgement and it does not prove that a doctoral opportunity is current. Crossref metadata enrichment can help you understand publication metadata and DOI connections, but a publisher or institution page remains the better source for a current research claim.

The ORCID guide to getting started and its explanation of source information show why researcher records can contain useful context while still needing a check against the university's current page. A profile can be self-managed or updated at a different time from an institutional role page.

Record the date you checked, not an invented freshness date

The check date tells you when you last looked. It does not prove that the source will remain current. Keep the exact date in the private record and use a review reminder based on the risk of the claim.

Time-sensitive claims include closing dates, funding amounts, eligibility, visa rules, fee treatment, and application routes. Recheck them before you act, even if your earlier check was recent. Stable research context, such as a group's stated methods, may need a longer interval, but it can still change.

Do not change a publication date to make a page look fresh. If a source shows an update date, record it as the source's date. If it does not, do not invent one. Your record can show that you checked the page on a particular day without suggesting that the opportunity itself was newly published.

Look for archive and closure signals

Search pages often keep old items visible. A call may be labelled closed, filled, archived, expired, or no longer accepting applications. A vacancy may remain online after recruitment. A programme page may describe a recurring route without saying whether the current intake is open.

When you see a closure signal, move the record to monitor or closed and save the official page that shows the status. If there is no clear signal, follow the application link and check whether the form accepts submissions or whether the page directs you to a new call. Do not infer current availability simply because the page still loads.

Set removal rules

Your system needs permission to get smaller. Close or remove a lead when the official source says that the route is closed, the host confirms that recruitment has ended, the route is outside your scope, or the unresolved question is no longer worth your effort.

Keep a short closed note with the reason and source. This stops you from reopening the same lead through a new search result and helps you recognise patterns in your search. A closed record is not a failure. It is a decision that protects your time.

Treat conflicts as work items

If two official pages disagree, do not choose one silently. Record the exact conflict, the pages, the dates shown, and the organisation that can resolve it. Ask a focused question and pause the decision that depends on the answer.

For example, a central admissions page may state one language route while a programme page lists another. A funder page may describe support generally while a provider page gives a different fee category. The right response is not to average the two statements. It is to identify which page controls the applicant's route and ask the named office if the relationship remains unclear.

Use a decision log and a comparison queue

The opportunity record says what the route is. The decision log says why you changed its state. Keeping both prevents your list from becoming a set of unexplained colours.

Write one sentence when a state changes

Whenever a lead moves, write one sentence that links the change to evidence. Examples:

• “Moved to verification because the employer page names the project and application portal, but funding coverage is not yet stated.”
• “Moved to monitor because the provider says the next recruitment call will be published separately and no current call is open.”
• “Paused because the programme page asks for a research background that I need the admissions office to assess.”
• “Closed because the official vacancy page says recruitment has ended.”

The sentence does not need to be long. It should make the reason visible to you a week later.

Keep an active queue small

An active queue is the set of leads that deserve a decision or a concrete action now. Set a personal limit that you can actually read and maintain. If the queue grows beyond that limit, move weaker items to monitor or close them after a quick evidence check.

There is no universal correct number. The right size depends on your time, document readiness, and the complexity of the routes. The principle is that an active lead should receive attention before its next meaningful stage passes.

Create a comparison queue after verification

Do not compare every discovered result. Comparison is useful after you have checked enough evidence to know that the items are real candidates. Add a lead to the comparison queue only when you have its identity, route, source, current status, funding record, and fit note.

The queue can link to the separate guide on comparing PhD opportunities. P11 owns the organisation that makes comparison possible. P12 owns the detailed choice between verified options. P01's discovery guide covers the search routes themselves, while P04's verification guide focuses on testing one opportunity before you invest in an application.

Use reasons instead of scores when evidence is thin

A score can hide the difference between “strong fit” and “unknown”. Early in the search, write reasons. If you later use a simple rating, keep the evidence notes beside it and include an “unknown” state. Do not reward a lead merely because it has a polished page or a large institution name.

The DORA declaration is a useful reminder that research assessment should not be reduced to one metric. The same habit applies to your search. A university ranking, a prominent funder, or a high number of publications may be context, but none of them answers all the questions about a particular route.

Manage documents and references without mixing versions

The search system should help you decide where to apply. It should not become the only home for every application file.

Keep one source record and separate application records

Create a stable opportunity record with the source facts. When you choose to apply, create an application record linked to it. The application record can hold your proposal, statement, CV version, writing sample, referee tasks, portal screenshots, and submission confirmation.

This separation solves two common problems. First, you can update a source fact without changing an old application version. Second, you can see which document was prepared for which route. A generic CV can be a starting point, but the submitted version should be identified by the application and date.

Name documents for retrieval

Use a simple naming pattern that includes the host, route, document type, and version. For example: HostProjectstatementv2 is more useful than finalstatement_new. If a portal asks for a specific file name or format, follow the portal instruction.

Do not keep the only copy in a browser download folder. Store the file in a place you can access when the portal is open, and keep a submission confirmation with the application record. Follow the university's privacy and file rules when storing identity documents.

References need their own status

A referee may need to upload a letter directly, complete a form, or respond to a request from the institution. Your search record should say that a reference is required and whether the current route states a deadline or submission stage. Your application record should track the person, request date, response status, and any reminder you choose to send.

Do not assume that a referee who supported one application can reuse the same letter for another route. The programme may require a different form, relationship, or direct submission. Read the current instructions.

Protect personal and source information

Opportunity records can contain personal documents, contact details, and links to restricted portals. Keep access limited and use the storage and privacy practices appropriate to your situation. Do not paste identity numbers, full financial records, or private referee correspondence into a public note.

The source-backed principle applies to personal evidence too. Record what a document proves and for which process. A transcript can support an academic requirement. It does not automatically prove a funding condition or an immigration requirement.

Use AcademicWings as one part of the system

AcademicWings is useful when it helps you move from a broad search to a clearer next decision. Start with the get-started page to create the search profile that matches your field, route, and timing. Keep your own source record alongside any product view so that the controlling page remains visible.

You can browse public opportunity records through opportunities, then follow the direct source and check the details that matter to your case. Use the guides to support practical application work and the methodology page to understand how source quality and uncertainty should be handled. AcademicWings does not replace the university, employer, funder, or government authority. It should not be described as verifying a route, saving an application, recommending a programme, or applying on your behalf unless the current product explicitly provides that action.

The most useful product habit is to keep the system honest. If a public record helps you discover a lead, capture the direct source. If a page does not state funding or eligibility, keep that gap in your notes. If your fit is uncertain, mark the question instead of converting it into a product label. A search profile can organise your starting point, but your decision still depends on current evidence and your own circumstances.

Your first week: build a system you will keep using

You do not need to document every historical link before you begin. Set aside one short session for the scope statement, then capture a small set of leads that represent the routes you are considering.

During the first session, write the research problem, degree boundary, geography where it matters, and timing window. Choose one place for the records. Add the fields in the template and create the states. Keep the system simple enough to update from a phone or a browser tab.

During the next session, capture a few leads from official discovery pages and university or funder sources. Give each one an identity, source type, five core claims, fit note, and initial state. Move only the strongest leads into the verification queue.

In the following sessions, verify one lead at a time. Follow the source hierarchy. Record what is stated, what is not stated, and who owns the next answer. Close duplicates and old results as you go. Do not wait for a perfect spreadsheet before making a sensible decision.

At the end of the week, look at the active queue. If it contains items with no direct source, move them back to discovery. If it contains items with unresolved high-risk claims, pause them and write the exact question. If it contains verified routes that fit your scope, hand them to the application workflow.

Run the search in short, repeatable cycles

A search system becomes useful when it has a rhythm. Without a rhythm, you either check everything every day or leave the list untouched until it is no longer trustworthy. Use short cycles with a clear purpose. One cycle can be for discovery, another for verification, and another for application preparation.

The discovery cycle

In a discovery cycle, search only within the scope you wrote. Use a small set of query combinations and visit the official pages behind the results. Capture new leads quickly and close duplicates. Do not try to verify every detail at the same time. The output should be a manageable group of records in the discovered state, each with a reason it fits your scope.

At the end of the cycle, look for patterns. Are the same hosts appearing? Are most results programme pages rather than vacancies? Are you finding projects with no funding details? This tells you how to adjust your next query. If your search keeps returning the same broad directories, add a research method, group, funder, or provider term instead of adding more generic degree terms.

The verification cycle

In a verification cycle, do not add new leads unless they are needed to resolve a gap. Choose the oldest or most promising records in the verification queue. Open the controlling page, follow the linked instructions, and complete the evidence block. If the source is no longer current, move the record to monitor or closed and record why.

Verification can feel slower than discovery because it asks you to read carefully. That is the point. A short queue of well-understood leads is more useful than a large queue of attractive titles. If you discover a second route while checking the first, capture it as a separate lead and return to the current record. Do not let one page turn into an unplanned browsing session.

The decision cycle

In a decision cycle, review only records that have enough evidence for an action. You may decide to apply, ask a question, monitor a future call, pause for a document, or close the lead. Write the reason and hand the record to the next part of your workflow.

If your decision is “ask”, write the question before you write the email. If it is “apply”, open the application record and list the required steps. If it is “monitor”, save the exact page to revisit and state what event would make the route active. A monitor item without a trigger is just an old bookmark.

Review the system itself

After several cycles, inspect the system rather than only the opportunities. Remove fields you never use. Add a field when the same missing answer causes repeated confusion. Merge records that refer to one route. Check whether your states still describe the decisions you make.

This small maintenance step matters because a search system can become a new source of noise. More columns, labels, and colour codes do not create better evidence. Keep the structure that helps you answer a question, and remove the structure that only records activity.

When the source is difficult to read or access

Official information is not always presented in a convenient format. A university may split the requirements between a programme page, a graduate-school policy, and an application portal. A funder may publish a call as a PDF with separate terms. A page may load through a search tool but not show the full text until you select a route.

Do not treat inconvenience as permission to guess. Start by recording the page you could access, the part it answers, and the missing link. Follow the official navigation and look for the named office or document. If the page says to use a portal, open the portal and record its instructions separately. If a PDF is the controlling document, save its title and the relevant page number.

If a source is temporarily unavailable, keep the lead in the verification queue with a note that the source could not be checked. Set a review action rather than moving it to active. A cached search snippet, social post, or third-party summary can tell you where to look when the page returns, but it cannot close the evidence gap.

When two versions of an official document are available, use the version linked from the current route and record the issue or publication details shown by the organisation. Do not assume that the file with the newest-looking filename controls the application. If the relationship between versions is unclear, ask the named office and keep the question in the record.

Keep source-backed writing honest in your own notes

The way you write a record affects the way you later write an application. Strong notes use plain, careful verbs: “the page states”, “the call asks for”, “the provider describes”, “my transcript shows”, and “I still need to confirm”. Weak notes use promises: “this will cover”, “they accept everyone”, “the supervisor will fund”, or “this guarantees a place”.

Use the same discipline when you write a research-fit paragraph. Cite the group's current page or publication for what it does, then explain what you hope to learn or contribute. Do not present a publication list as proof that a supervisor is recruiting. Do not turn a programme's research theme into a promise that your proposal will be accepted.

If you later share a shortlist with a referee, supervisor, or adviser, the source and check date make your reasoning easier to inspect. They also make corrections easier. Someone can challenge a specific claim without having to reconstruct your entire search history.

The Google helpful content guidance is aimed at publishers, but its people-first principle is useful here too: make each note useful for the decision a real person must make. Do not produce a long summary that hides the one unresolved question. A short record that tells you what controls the next step is better than a polished paragraph with no evidence.

The Source-Backed Search Truth Check

Before you call a lead an active PhD opportunity, check the record from top to bottom.

Identity and ownership

• Is the title exact enough to find the route again?
• Is the host, department, employer, or consortium named?
• Is the funder or provider recorded separately when relevant?
• Do you know which organisation receives the application?
• Is the current controlling URL saved, not only a directory or search result?

Route and status

• Is this a vacancy, programme, network, award, fellowship, or research lead?
• Does the source show that a current candidate action is available?
• Is a separate admissions, nomination, supervisor, or funding step stated?
• Have you separated active, monitor, paused, and closed items?

Evidence

• Did you record what the source says about eligibility, funding, documents, and timing?
• Did you mark missing information as “not stated” instead of guessing?
• Are important claims supported by the organisation that controls them?
• Did you save linked terms, forms, or programme instructions?
• Is the check date recorded and is the next review appropriate to the risk?

Fit and decision

• Is your fit note clearly separate from the institution's wording?
• Have you named both your evidence and your possible gaps?
• Does the next action have one owner?
• Can you explain why the lead is active, monitored, paused, or closed?
• If you apply, is there a separate application record for your documents and references?

If several answers are “no”, the record is not ready for an application decision. That is not a problem. The check tells you exactly what to do next.

Official sources for this guide

The following official pages provide the source examples and research context used in this guide. They are not a substitute for the current controlling page for your own opportunity.

• DAAD: finding a PhD position in Germany, a discovery workflow example.
• EURAXESS Jobs, a discovery service for research vacancies.
• MSCA Doctoral Networks, an example of a network structure whose project documents control individual conditions.
• MSCA Postdoctoral Fellowships, an example of a host-led fellowship call.
• UKRI doctoral training investment framework, an example of provider and partnership funding structure.
• UKRI studentships and doctoral training, an example of how provider pages connect students to studentship routes.
• Directory of Open Access Journals, a journal discovery and transparency resource with limits on what inclusion proves.
• Crossref metadata enrichment, background on publication metadata and DOI information.
• ORCID: getting started, a guide to researcher records.
• DORA: read the declaration, a reminder not to reduce research assessment to one metric.
• Google helpful content guidance, a people-first publishing standard.

The source layer helps you understand the workflow. It does not decide whether a specific opportunity is open, funded, suitable, or available to you. For that decision, return to the current university, employer, funder, provider, or government page and record the evidence that controls your next step.