Philippines staffing guide · Published
Write a research source log for delegated market scanning
A summary hides where the claims came from. A source log lets an assistant scan a market while keeping each claim, date, and access path reviewable.
The short answer
Record the question before the search, capture a source and access date for every claim, grade the evidence, and keep open doubts visible. The log turns delegated scanning into something a decision owner can check.
Why a source log beats a summary
A market summary is easy to read and hard to verify. It compresses many claims into a short narrative, and the compression hides which parts came from a regulator and which came from a vendor blog.
A source log keeps the claim, the source, the date, and the access path together. The summary can still exist, but it should be a view of the log rather than a replacement.
A log also changes the assistant's behavior. When every claim needs a source, the assistant stops filling gaps with plausible language. The work becomes slower at first and more useful after the first week.
The reviewer can spot a weak claim in seconds instead of rereading the whole document. That speed is the reason to keep the log even when the final deck looks polished.
Record the question before the search
Write the decision the research is meant to support, the population or market being studied, and the time window. A question such as is this market growing is too broad to answer honestly in a short scan.
A better question names the segment, the comparison, and the decision it informs. The log should open with that question so a later reader knows what the scan was trying to do.
Add the exclusions before collection starts. State which regions, company sizes, or time periods are out of scope, and why. Without exclusions, the assistant will drift toward whatever sources are easiest to find.
The reviewer then cannot tell whether a missing segment was deliberately excluded or simply overlooked. Exclusions are part of the method, not an apology at the end.
Capture source, date, and access path
For each claim, record the source name, the URL or document identifier, the publication date, and the date it was accessed. Record the exact table or section when a source contains many numbers.
A link alone is not enough, because pages move and content changes. The access path tells the next reader how to find the same statement again.
Note whether the source is primary or secondary. A regulator's dataset is primary, while a news article describing it is secondary. Both can be useful, but they carry different weight.
If a secondary source cites a number, follow the citation to the primary source when it is available. If it is not available, say so in the log instead of presenting the number as confirmed.
Grade claims by evidence type
Use a small scale: verified from a primary source, supported by a credible secondary source, or unverified. Do not use the same label for a regulator's table and an opinion in a trade publication.
The scale should be simple enough to apply consistently and strict enough to separate fact from inference. Write the definition beside the scale so a new reviewer applies it the same way.
Keep the assistant's own interpretation in a separate field. A claim describes what the source says, while an implication describes what the team thinks it means. Mixing the two makes the log look more certain than it is.
When an inference depends on a missing fact, label it as an open question. The decision owner can then choose whether to spend more time or proceed with the uncertainty.
Review the log for gaps
Before the log goes to the owner, check that every headline claim has a source and every source has an access date. Look for claims that appear only once and carry a lot of weight, because those are the most fragile.
Check whether the sources represent more than one perspective. A scan built from vendor material will describe a market differently from one built from regulator data and customer interviews.
Have a second reviewer sample five claims and try to reproduce them from the recorded path. If the reviewer cannot find the statement, the log is not complete.
Record the reproduction result and fix the path or remove the claim. This check takes less time than a full reread and catches the errors that damage trust most.
Hand the log to a decision owner
Present the log with a short cover note: the question, the strongest findings, the weakest evidence, and the decisions that remain open. Do not bury the uncertainty in an appendix.
The owner needs to see where the evidence is thin before making a commitment. A clear limitation is more useful than a confident conclusion that cannot be defended.
Keep the log after the decision. It becomes the record of what was known at the time, which is valuable when the market changes or the decision is revisited.
Update it with a new date rather than overwriting the old version. The history shows how the team's understanding changed, and it stops a later reader from treating an old estimate as current fact.
Keep the log useful after the decision
A source log can become a graveyard of links if nobody maintains it. Add a simple status for each source: current, superseded, or unavailable.
When a regulator updates a page, note the change against the old claim instead of silently replacing it. The status tells a later reader whether the evidence still supports the original conclusion.
At the next scan, start from the open questions and unavailable sources rather than from a blank page. Recheck the claims that carry the most weight and leave stable findings alone.
This pattern keeps the work cumulative and prevents the team from paying for the same research twice. A maintained log becomes an asset, not a filing chore.
Questions to copy for the sales call
- "What decision is this scan meant to support?"
- "Which claims are verified from a primary source?"
- "Where is the weakest evidence, and who needs to see it?"
- "Can a second reviewer reproduce five claims from the log?"
Sources
- U.S. Census Bureau: Data Tools and Apps: Primary demographic and business datasets that can anchor a market scan.
- BLS Handbook of Methods: Explains how major labor and price series are constructed, which helps a reader judge comparability.
- GAO: Assessing Data Reliability: A structured approach to testing whether data is complete and reliable for a decision.
Common questions
How many sources should a short scan use?
Enough to cover each main claim with a primary or credible secondary source. A short scan often needs five to ten sources, but relevance matters more than count.
Can the assistant use news articles?
Yes, as secondary evidence. Follow citations to primary data when possible and label the claim as secondary when the original is unavailable.
Should the log include quotes?
Short exact quotes help when wording matters. Otherwise paraphrase and link the source, keeping the quote inside fair-use limits.
Who owns the final judgment?
The internal decision owner. The assistant assembles and grades evidence; the owner decides what the evidence means for the business.