Use Factual Research to Change the Series Question, Not Fill the Deck
Use Factual Research to Change the Series Question, Not Fill the Deck
You can usually spot the research slide from the back of a room. It's the densest page in the deck, three columns of type too small to read, and if you deleted it nothing else in the deck would move.
That's the tell. The research is real, the citations are real, and the page is inert.
The test that fixes this is blunt. Research earns its place in a factual-series pitch when it changes one of three things: what the series will ask, who or what it will follow, or what it can honestly claim to establish. A finding that changes none of those isn't bad research. It's background, and background belongs in the file you keep for yourself.
This is a harder test than it sounds, because almost any finding is compatible with the pitch you already have. A series about a tool library and a finding about tool libraries will always get along. Compatibility is cheap. The question is whether the finding moves a decision.
A note on what follows. I don't have an on-the-record account of a real production changing its series question this way, and I'm not going to describe one as though I did. The example below is invented — a fictional tool-library series with a small, fixed set of research materials — and it's here to show the mechanics. Take it as a method, not a reported case.
Write the assumption hidden inside the series question
Start with the question you're already carrying around. Say it's this: Why do borrowers fail to return what they borrow?
Read it again and it's doing a lot of quiet work. "Fail" carries a verdict before any evidence arrives. The question assumes three things at once: that non-return happens, that it happens often enough to sustain a series, and that the reason sits on the borrower's side of the transaction. A question that smuggles in its own answer is hard to research honestly, because every compatible finding starts to feel like progress.
So the first job is to unpack it into the claims it actually needs. For the tool library there are two, and they're separable:
- Tools leave and don't come back, in quantities worth building a series around.
- The explanation for that belongs to the borrower — forgetting, carelessness, deciding to keep it.
Those are different claims. A piece of evidence can support the first and say nothing about the second, and the temptation is to read support for one as support for both.
There's a third thing to check, and it's practical rather than intellectual: which version of the question can you film? A blame question needs the blamed party — a borrower who didn't bring the tool back, sitting in a chair, willing to explain. That's a fragile spine for a series. If the proposition rests on people showing up to account for themselves, it's hostage to casting you may never win.
Write the assumption down. Now you have something to compare findings against, which is the only way to notice when a finding has changed your direction rather than confirmed it.
Translate a finding into a bounded editorial change
For every candidate finding, keep four short lines:
- Where it comes from.
- What it covers — the real scope.
- What it establishes.
- What you'll change because of it.
Line four is yours. A finding doesn't make an editorial decision; it gives you something to decide with. If you can't name the finding that caused a change, the change probably came from a preference, and a preference wearing a citation is worse than an honest hunch.
Here's the fictional pack, item by item.
The checkout log. A handwritten desk log for one cabinet of hand tools, with a column for the date a tool came back. Its scope is narrow: one cabinet, one logbook, only the checkouts that happened at the desk. What it establishes is thin but genuine — that these tools went out on these dates under these names. What it does not establish is that any particular tool failed to come home, because a blank in the returned column is a blank, not a fact. The change: the log stops being your evidence about borrower behavior. It becomes evidence about the record itself.
The volunteer's account. One volunteer describing how returns get recorded — often when nobody is at the desk, written in later, sometimes not written in at all. This is testimony about a process from a single person. It doesn't tell you how often the column goes unfilled. What it does is disqualify the blanks you were just looking at, and — more usefully — it names a moment. A tool comes back over a counter. Somebody writes something down, or doesn't. The change: the desk and its logbook become a place the series could actually watch.
The borrower explanations. A small set of them, gathered loosely: a few people who forgot, one who lent the tool onward, one who assumed there was no deadline. Self-selected, not a sample. What they establish is the shape of the problem — "didn't return it" turns out to cover several behaviors that only look alike from the library's side of the desk. The change: borrowing failure stops working as a single category, and the question has to be wide enough to hold more than one of these.
Three findings, three changes. None of them settled the original question, and none supplied the frequency it needed to sustain a series.
Note what the changes are not. Deciding to film the return counter is not the same as having permission to film there. A proposed editorial change stays a proposal until access exists, and if the deck blurs that line, you'll spend the meeting negotiating about something you described as settled.
One more thing, about confirmation. A finding that supports your existing question can still earn its place — if the decision was genuinely in doubt. If you were one phone call away from dropping the premise and the log settles it, that's a change. If nobody was ever going to drop it, the confirming finding is decoration.
Let contradictory material reshape the inquiry
The interesting thing about the tool-library pack is that its pieces disagree, and they disagree unevenly.
The blank return column points one way: tools not coming back. The volunteer's account points another: the record doesn't reliably show returns at all. The instinct is to weigh them — two sources against one, or the document against the recollection. Don't. They aren't peers, because they aren't answering the same question. The volunteer is describing how the record was made. A record can't outvote the person describing its making. That's the difference between a real conflict of evidence and a tally of opinions that happen to share a subject.
When unlike records disagree, you have three honest options, and choosing among them is the work:
- Change the question. If the disagreement can't be resolved with what you have, maybe the original question demanded a certainty the material can't supply. A blame question can't survive an unreliable record. A question about the process around the record can.
- Report more. Some contradictions are answerable. A few weeks of sitting at the desk and writing down every return as it happens would tell you something the logbook can't. That's a reporting plan, not a finding, and it should be labeled as one.
- Block a claim. Sometimes the disagreement means you don't get to say the thing you wanted to say. "Borrowers don't return tools" is not available — not until something else arrives.
And watch for the failure that looks like rigor: deciding the borrowers were fine and the library's records are the real story. That's a new verdict wearing the old one's clothes. You've removed one explanation by fiat and installed its opposite. The volunteer's account limits the log; it doesn't exonerate anybody. Nobody has counted returns yet.
The borrower explanations sit off to the side of this. They don't resolve the log question. They widen the category, which is its own kind of reshaping, and they come from people who chose to talk to you — so they show you the range of reasons an item might not come back, not the frequency of any of them.
Notice what hasn't happened. Nothing is settled. The competition between "they don't bring things back" and "nobody writes down what comes back" is still open, and the size of the real problem is still unknown. That isn't a failure of the research. It's the reason there's a series. If the pack had closed the question, you'd have an answer, not a run of episodes.
Replace the research dump with the changed proposal
Now the deck.
The question becomes: What happens between a tool leaving the library and becoming available again?
Plainer, and less exciting than what you started with. It's also filmable, because it's about a place and a sequence rather than somebody's character. A tool goes out. Time passes. A tool comes back over a counter. Someone writes or doesn't. The tool lands on a shelf, or it doesn't. That's a process you can stand next to with a camera, and it's a process where the interesting thing — whether the record matches the world — happens in the open.
The person changes too. The old question sent you hunting for a borrower who kept a drill. The new one puts you beside whoever is working the desk.
Into the deck go two things and no more. The volunteer's account, attributed as one volunteer's description of how returns get logged, with its limits stated plainly. The borrower explanations, described as a small self-selected set, offered as a range rather than a rate. That's enough for a reader to see why the question moved.
Out of the deck goes the page of general figures about how much money shared tools save households, or how much landfill they divert. It isn't wrong. It's just that nobody in the room was going to make a different decision because of it. The value of tool libraries was never the disputed part. Put that page in the research file with everything else, where it can be useful later and won't be mistaken for a reason.
Also out: any sentence that makes a plan sound like a result. "We will film the return counter" can sit in the deck as an intention, flagged as unconfirmed access. "We watched a tool come back and go unlogged" doesn't belong anywhere yet, because it hasn't happened.
Keep the trail outside the main argument — a separate document listing what you have, what it covers, and what you still need. The deck shows the decision. The trail shows your homework. Mixing them is how a pitch turns into a reading assignment.
That's the whole movement: research in, a decision out, and the parts that changed no decision staying in the file. If the deck is doing its job, a reader can point at one finding and one sentence and see exactly how they connect — and can point at the empty space and see what's still being negotiated.
For the tool library, the series now asks what happens between checkout and the shelf, and the change that got it there is small and traceable: one volunteer's account of how returns get recorded, which turned a blank column from evidence into a question. What stays open is whether returns can be counted at all — whether the desk will let anyone sit beside it for a few weeks, and whether the tools that never came back are a real group or a shadow cast by an empty column. That's a reporting problem now, not a slide, which is the better place for it.
Frequently asked questions
What test should decide whether a factual research finding earns a place in a series pitch?
It earns its place if it changes what the series will ask, who or what it will follow, or what it can honestly claim to establish. A finding that changes none of those is background for the research file, not a slide.
Why is compatibility between a finding and a pitch not enough?
Almost any finding on the same subject is compatible with the pitch you already have. Compatibility is cheap. The question is whether the finding moves a decision. A confirming finding can earn its place only if the decision was genuinely in doubt; otherwise it is decoration.
How should the fictional blank return column be treated?
A blank in the returned column is a blank, not a fact that a tool failed to come home. The volunteer's account can disqualify the blanks and limit the log, but it does not exonerate borrowers or establish frequency. Nobody has counted returns yet.
What changed in the tool-library series after the research pack?
The question moved from why borrowers fail to return to what happens between a tool leaving the library and becoming available again. The focus moves to the desk and the process. The deck includes the volunteer's account with limits and borrower explanations as a range, not a rate; general savings or landfill figures go to the research file.
When unlike records disagree, what honest options are available?
Change the question, report more, or block a claim. Do not weigh them as peers when they answer different questions; a record cannot outvote the person describing its making. Avoid replacing one verdict with its opposite, such as deciding borrowers were fine and records are the real story.