Case Study 40.2 — The Curation Session: Building a Final Portfolio on the Kitchen Table, Start to Finish
"To collect photographs is to collect the world." — attributed to Susan Sontag
The brief and the constraints
Case Study 40.1 took a finished masterwork apart to see how selection and sequence make meaning. This one does the opposite: it builds a portfolio from scratch, in real time, on a kitchen table — the same kitchen window location that has anchored so many shoots in this book, now repurposed as a curation studio. There is no new photography here. The "shoot" is the editing session, because in the capstone the editing is the production. You will watch the entire process — cull, sequence, present — happen as a sequence of decisions, with a Described Photograph of the work at each stage and a note on what changed and why.
The brief, then, is the same one the Portfolio Checkpoint set you: take a full pool of candidate images accumulated across the book and turn it into a finished, sequenced, presented portfolio of 20–30 frames. The constraints are deliberately humble and universal, so any reader — phone-only included — can run the identical session:
- Location: a kitchen table by a window. Good soft light to view prints by (Chapter 5), a flat surface big enough to lay everything out, and room to move pieces around with your hands.
- Materials: the full candidate pool as small prints or printed thumbnails (four to a page is fine), a pen, sticky notes, and a phone or notebook to record decisions. Physical, not on a screen — the hand is smarter than the cursor at this.
- Time: one long session for the cull and sequence, then a second session on a different day to finalize (the sleep-on-it pass of Figure 40.2 is non-negotiable).
- No new frames. If you are tempted to go shoot something to "fill a gap," resist — work with what you have. The gap is information about your range, not a problem to paper over today.
We will follow one representative candidate pool through the whole session. The pool and its images are a constructed teaching example, but the process is exactly the one to run on your own material.
⚙️ Settings Box — Viewing conditions for a curation session (not a camera setting):
Variable Setting Why Light to view by Soft, neutral daylight (north window) or a neutral lamp ~5000K Color-cast light lies about your prints (Chapter 27); neutral light tells the truth Surface Large, flat, matte, uncluttered You must see all candidates at once and move them freely Print size Small (4×6 or thumbnails, 4-up) Cheap, fast, and forces you to judge the image not the detail Distance Stand back; view the whole layout from a step away Sequencing is about the set, seen as a whole, not one frame up close Company Solo for the cull; one trusted eye for the shortlist Solo keeps the cull honest; an outside eye breaks the favorite-child trap
The candidate pool
Here is the raw material we will build from — a constructed but realistic pool of 34 candidate images, the kind a reader finishes this book holding. Spread on the table, it looks like this:
FIGURE CS40.4 — "The pool: 34 candidates on the kitchen table" [constructed teaching example]
THE FRAME Thirty-four small prints laid in a loose grid on a pale wooden table by a window. Reading
across the clusters that emerge at a glance:
• 9 golden-hour landscapes (six of them the same valley overlook from Chapter 16, barely different)
• 4 window-lit portraits (Chapter 13) — two strong, two competent
• 3 red-door frames (the Chapters 7, 26, 28 versions — color, RAW, B&W)
• 5 street frames from the market (Chapter 17) — one decisive moment, four ordinary
• 2 long-exposure night/neon shots (Chapter 21) — one striking, one muddy
• 1 motion-blur panning shot of a cyclist (Chapter 20)
• 3 macro frames (Chapter 19) — one excellent, two redundant
• 2 black-and-white tonal studies (Chapter 8)
• 1 high-key still life (Chapter 14)
• 2 "darlings" — technically flawed images kept for personal reasons
• 2 establishing/sense-of-place frames (Chapter 24)
THE LIGHT Soft, even window light across the whole table — flaws and strengths equally visible, nothing
flattered or hidden.
THE MOMENT Before the first cut: the pile at its most overwhelming, every image still a candidate.
THE EFFECT The eye immediately spots the redundancy (six near-identical valleys) and the standouts (the
one decisive market frame glows even at thumbnail size). The clusters announce the range — and
the gaps.
THE LESSON Spread everything out and the portfolio half-curates itself: redundancy and strength are
visible at a glance only when you can see the whole pool at once.
Thirty-four candidates. The target is 20–30, but the real target is "every frame earns its place," which will land us lower than the maximum. Let's run the session.
Phase 1 — The fast pass (gut, seconds each)
First pass is pure instinct, exactly as Figure 40.2 prescribes: pick up each print, give it a fast yes / maybe / no, set it in one of three piles, and do not deliberate. The goal is to clear the obvious noise quickly so the real work happens on a smaller field.
The thinking, sped up: the six near-identical valley overlooks — one obvious yes (the best light), the rest straight to maybe or no; we cannot keep six of the same view. The muddy night shot: no. The two redundant macros: no, keep the excellent one as yes. The four ordinary market frames: no; the decisive one is an instant yes. The two darlings: maybe (we know what's coming, but we don't cut them on the gut pass — that decision deserves the formal screen). The competent-but-anonymous second-string portraits: maybe.
FIGURE CS40.5 — "After the fast pass: three piles" [constructed teaching example]
THE FRAME The 34 prints now sorted into three stacks. YES (clear keepers): 11. MAYBE (uncertain): 12.
NO (gone): 11. The NO pile is pushed firmly to the table's edge — out of sight, out of the
decision.
THE EFFECT The field has halved. The YES pile already looks like the spine of a portfolio; the MAYBE
pile is where the real judgment now concentrates.
THE LESSON The fast pass is not the portfolio — it is *clearing the field* so the five-question screen
runs on 23 images instead of 34. Speed first, rigor second.
What changed and why: nothing precious was cut yet — only obvious redundancy and clear technical failures. The darlings survived to the next pass on purpose; cutting them on instinct would deny them the fair, explicit test that makes the cut defensible rather than merely felt.
Phase 2 — The five-question screen (one image at a time)
Now the YES and MAYBE piles (23 images) go through the screen of Figure 40.1, one at a time, in writing on sticky notes. This is slow and it is the heart of the session. Two worked examples show the screen doing its job.
The first darling — a soft, slightly soft-focus portrait of the photographer's grandmother.
- Technically sound? Marginal — focus is slightly soft on the eyes. Strong on its own, no story? This is the one that hurts: cover the fact that it's grandmother, and it is a pleasant but unremarkable portrait, a little soft. Range or depth? Redundant — there are two stronger window-lit portraits already. Yours? Not distinctively. Name test? Hesitation. → CUT. The note reads: "Keep forever, in a frame, on the wall at home. Not in the portfolio. The love is real and invisible."
The one decisive market frame — a vendor mid-gesture, hand raised, caught at the exact instant (Chapter 10, 17).
- Technically sound? Yes — sharp where it counts, exposure holds. Strong on its own? Yes — it works with zero backstory; a stranger feels the moment. Range or depth? Yes — it's the only true decisive-moment frame, widens the range. Yours? Yes — this is the kind of seeing that's becoming characteristic. Name test? Without hesitation. → KEEP, and flag it as a candidate opener — it announces a photographer who can see a moment.
Running the full 23 through the screen, the maybes resolve: both darlings cut; the two anonymous portraits cut (redundant with the two strong ones); one establishing shot cut (the second is stronger); the panning shot kept (it's the only motion frame — range); one more valley landscape cut. The contenders settle at about 18.
FIGURE CS40.6 — "The contenders: ~18 survive the screen" [constructed teaching example]
THE FRAME Eighteen prints in a tighter grid, each with a sticky note carrying its verdict and the one
reason. The cut pile at the table's edge has grown; notably it contains several GOOD images —
not failures, just images outranked by stronger ones.
THE EFFECT The set now reads as deliberate. Every frame has a defensible reason to be there, written
down. The range is visible: light, portrait, street, landscape, macro, night, motion, B&W.
THE LESSON The screen's power is that it makes you WRITE the reason — and "I can't write a reason" is
itself the verdict. Sentiment cannot survive being asked to justify itself in one sentence.
What changed and why: the darlings fell to criterion two (invisible backstory) and criterion three (redundancy), exactly as §40.1 predicted the most-beloved images would. Cutting them was the hardest moment of the session and the most important — it is the difference between a portfolio and a scrapbook.
Phase 3 — Kill redundancy and cut to the floor
Eighteen is within range, but the floor-raising pass remains. Group the survivors by cluster one more time and ask of each: raise the floor or sit at it? The weakest of the three remaining red-door frames is good but the other two are stronger and the trio is slightly redundant — cut to two. The second night shot is fine but the striking one carries the genre alone — actually, on reflection it adds a second mood (neon vs. light-trail), so it stays; redundancy is about sameness, not genre. One more landscape is competent but anonymous — cut. We land on 16.
Sixteen is below the 20–30 maximum, and that is the right answer here: a tight, all-strong set of sixteen beats a padded twenty-five. (If your honest pool supports more, keep more — the number serves the work, not the reverse.) The session pauses. This is where you sleep on it — the floor-raising verdict is made fresh, on a second day. On the second day, two images that felt borderline on day one now clearly belong, and one that felt safe reveals itself as filler and is cut. Final: 15 frames.
⚠️ Common Mistake: Curating in one sitting, in love with the pile. The single biggest failure of a curation session is doing it all at once, while emotionally warm to your own work — you keep too much because every cut stings in the moment. The fix is built into this session: the floor-raising pass happens on a different day, with fresh, colder eyes. An image you fought to keep on Monday is often the one you calmly cut on Thursday. If you skip the sleep-on-it pass, you will publish a portfolio one or two weak frames too long, and those frames will set the floor the whole set is judged by.
Phase 4 — Sequence on the table
Fifteen finalists. Now the order, built physically with the hands (§40.2). Clear the table and start with the two fixed points.
FIGURE CS40.7 — Sequencing layout on the table (top-down view of the prints)
OPEN ─► [decisive [golden [red-door [macro, [B&W tonal
market valley color single study,
moment] overlook] study] leaf] high contrast]
LOUD med med quiet med
│ │
▼ …then the rhythm rises and falls across the middle… ▼
[window [street [night [high-key [panning [establishing CLOSE ─►[quiet B&W
portrait] frame] neon] still life] cyclist] sense-of-place] red-door
LOUD med LOUD quiet LOUD med on the wall]
QUIET, resonant
Fixed first: OPENER = the decisive market moment (most characteristic — "I can see a moment").
Fixed last: CLOSER = the quiet B&W red door (the book's through-line; lingers; personal).
Then: thread the rest between them, breaking up runs of LOUD, inserting quiet frames to breathe.
The reasoning, spoken aloud at the table: the opener is the decisive market frame — not the single prettiest image, but the most characteristic, the one that says "this photographer sees moments" (primacy). The closer is the quiet black-and-white red door — the book's own through-line, a calm, resonant note to linger on (recency), and quietly personal in a way that earns the last word. Between them, walk the loudness map: the loud night-neon frame can't sit next to the loud window portrait, so a quiet still life goes between them to let the eye rest; the macro leaf and the wide valley are placed adjacent so the scale collapse (vast → tiny) does deliberate work; the two red-door frames are kept far apart so the motif recurs across the set rather than clumping (the Americans lesson from CS-01, at small scale).
Walking the sequence end to end as a stranger reveals one sag — a flat med-med-med stretch in the middle. The fix: promote the striking macro into the lull and cut the weaker of the two sense-of-place frames. Now 14 frames, and the sequence breathes from open to close.
FIGURE CS40.8 — "The finished sequence, walked as a viewer" [constructed teaching example]
THE FRAME Fourteen prints in a single row across the table, left to right, in final order. Standing back,
the row has a visible rhythm: it opens on a bright human moment, rises and falls through
landscape, light, street, and night, and settles to a quiet, dark, personal close.
THE EFFECT Walked in order it feels composed, not piled — each transition intentional, no flat stretch,
the strongest frames at the ends. The red door appearing early (color) and last (B&W) quietly
bookends the whole set, a motif that rewards the attentive viewer.
THE LESSON The same fourteen images in date order would feel like a folder; in *this* order they feel
like a statement. Order is the last and most underrated act of authorship.
Phase 4b — One sequencing decision, shown both ways
Sequencing is invisible when it works, so it is worth slowing down on a single decision to watch the reasoning. Late in the session, two frames competed for the third slot: the warm golden-hour valley landscape and the cold blue-hour cityscape. Both are strong; both are "medium" loudness. The question was not which is better — they are roughly equal — but which serves the sequence here. Lay the two candidate orders side by side:
FIGURE CS40.9 — Two orders for slots 1–4, compared [constructed teaching example]
ORDER A: [decisive market] → [golden valley] → [red-door color] → [macro leaf]
LOUD warm / med warm-red / med quiet
Effect: slots 2–3 are BOTH warm; the eye settles into a warm run, then drops to quiet.
Pleasant, but the warm-on-warm adjacency makes slot 3's red less surprising.
ORDER B: [decisive market] → [blue cityscape] → [red-door color] → [macro leaf]
LOUD cool / med warm-red / med quiet
Effect: the COOL cityscape in slot 2 makes the RED door in slot 3 jump (color contrast, Ch.7).
The warm/cool adjacency does work the warm-on-warm version can't.
Order B wins — not because the cityscape is a better photograph than the valley, but because placing a cool frame before the red door makes the red louder by contrast (the color-contrast principle of Chapter 7, now operating between frames). The valley landscape isn't cut; it moves to a later slot where its warmth pairs well with a different neighbor. This is the whole craft of sequencing in one decision: you are not ranking images, you are choosing adjacencies, and the same image is stronger or weaker depending entirely on what sits beside it. An hour of the session goes to exactly this kind of small, reversible swap — and it is the hour that turns a row of good prints into a sequence.
💡 Why It Works: Two frames of equal quality can produce unequal sequences, because sequencing is about relationships, not rankings. The cool-then-warm adjacency exploits simultaneous color contrast (the same mechanism that makes a red door pop against a grey wall within a frame, Chapter 7) at the scale of the set. This is why you sequence with prints you can physically swap: the right order is found by trying adjacencies, not by reasoning about images in isolation.
The edit pass: presentation
The set is curated and sequenced; the production finishes with presentation (§40.3). Three quick moves at the same table:
- Print check. The 4×6 work prints already revealed the truth (the soft grandmother portrait's softness was obvious on paper; the decisive market frame glowed). For the final, order a small set of better prints of the finalists to confirm color under neutral light — soft-proofing against the lab profile so the calibrated edit (Chapter 27) survives to paper.
- Web gallery. Translate the table sequence to the screen: the grid order matches the row, opener top-left, generous whitespace, mobile-first, alt text written for every frame (a Described Photograph in miniature — the skill this whole book trained).
- Artist statement. Three short paragraphs (Figure 40.4): what — "I photograph the ordinary edges of my own neighborhood, in the hour the light goes strange"; why — "I'm after the moment a familiar place stops being invisible"; how — "made on foot, on whatever camera was in my pocket, over the months of one book." Read aloud, banned words cut. It points at the work and steps back.
Discussion questions
- The session cut two "darlings" that the photographer loved. Walk through exactly which curation criteria cut each, and describe a darling of your own and which criterion would cut it.
- The final set landed at 14 — below the 20–30 maximum. When is a smaller, tighter portfolio the right answer, and when does going small mean you simply haven't shot enough range yet?
- The opener was the most characteristic frame, not the single best, and the best was held for elsewhere. Why might "most characteristic" beat "most beautiful" for the first slot?
- The two red-door frames were deliberately placed far apart so the motif would recur. How is this the same principle as Robert Frank's recurring flags and jukeboxes (CS-01), and how could you use a recurring motif in your own sequence?
- The floor-raising verdict was deliberately made on a second day. Describe a time your judgment of your own work changed overnight — and why building that delay into the process matters.
Your turn
Run this exact session on your own pool, on your own kitchen table. Print or thumbnail every candidate, spread them out by a window, and walk the four passes: fast yes/maybe/no, the five-question screen (write the reason on a sticky note for every survivor), kill redundancy, then stop and sleep on it before the final cut. The next day, sequence physically — fix your opener and closer first, then thread between them by rhythm and rhyme, and walk it as a stranger to find the sag. Finish by printing a set, building the gallery, and writing the statement. The output is not an exercise; it is your finished portfolio. The session is the capstone.
Key takeaways
- The editing is the production. A curation session — cull, sequence, present — is itself a making process with phases, decisions, and a finished artifact; the "shoot" in the capstone is the editing.
- Spread everything out, physically. Redundancy and strength are visible at a glance only when the whole pool is in front of you; the hand sequences better than the cursor.
- Run coarse to fine, and sleep on it. Fast pass → five-question screen (write the reason) → kill redundancy → floor-raising cut on a different day. The most-beloved images fall to "no visible story" and "redundant," exactly as predicted — and cutting them is the point.
- The number serves the work. A tight set of 14 all-strong frames beats a padded 25; let the maximum be a ceiling, not a target.
- Order is the last act of authorship. Fix opener (most characteristic) and closer (lingering, resonant) first, thread by rhythm and rhyme, recur a motif across the set, and walk it as a stranger to kill the middle sag — then print, publish, and state it plainly.