“Flawless fruit across every parcel.” “Fruit expression on another level.” “Near perfect.” The loudest reaction to Clos Bertaud 2 is about one thing. Fruit is solved. It's the note everyone repeated.
On the narrow claim, the hype mostly holds up. Can the estate deliver clean, correctly ripened fruit, even off a difficult slope, even in a cold year? Mostly, yes. That problem really did get better.
But growing fruit and building structure are two different questions. Fruit just means getting the grapes right. Structure is the harder problem: setting acid and tannin at the right weight and grain, held in a balance that survives a real flight across multiple adjustments. The 42 sessions with 7 tasters surfaced a clear pattern. The estate is better than anything before it, but still not good at the structure that ships, and the gap is specific enough to predict.
The frame is right, the components are wrong
Clos Bertaud 2 gets the structural frame right and the components wrong. Every macro theme landed positive, but every micro theme landed negative.
In the deep dive, the tasters consistently worked on the structure. They fixed some of it in the glass and handed the rest to the blending bench.
Bertaud is good at the structural frame and bad at structural execution. Sort the feedback by theme and the divide is clean. Balance and composition (net +3), house fit and character (net +3), length and style (net +1). All decisions that shape the whole glass come in positive. Acid balance and definition (net −1), tannin grain and weight (net −3), the properties the estate has to hold sip by sip, come in negative.
Every macro theme landed positive. Every micro theme landed negative. On this sample, the split is clean with no single theme crossing the line.
Bertaud picks on-house varieties, lays out a balance that reads, and holds composition together across a dense blend. Then it stumbles on the individual components. A taster on the Sonnaille Collective flight said:
The fruit in the Sonnaille blend is hard to read, especially the bramble, black cherry, dried herb and iron layer at the back. Same with the tannin — the grain feels very generic and not defined.
Racking fixes one thing and leaves the rest
Racking fixes definition, the one thing you can ask for. Every structural problem stays where it started.
The round data shows the same split holding over time. Racking reliably clears one problem. Lacking definition, the issue where fruit sits too faint against the oak to read cleanly, falls from 3 mentions at Round 1 to zero by Round 3. On the Crête Noire flight you can watch it happen. The attack starts washed out, the mid-palate half-buried and hard to read. A taster asks for it to be sharper and match the house style, and the next round delivers it. This is the one problem tasters can fix by asking.
The structural problems don't move. Broken structure holds at 3 mentions through every round, three passes for zero improvement. On the same flight, a taster flagged that the Syrah in the Crête Noire blend reads as Grenache. That is what broken structure means, and three rounds of racking never fixed it.
Incorrect fruit stays at 1 the whole way. And on Sonnaille Collective, the problem grew. Muddled grain showed up at Round 2, flagged by two tasters independently, on a component the first pour didn't have.
Definition itself splits along the same line. Lift is the macro side of it, with one decision per blend, and the estate fixes it on request, which is why it clears by Round 3. Tannin grain and weight are the micro side, where individual components matter, and they don't move. Tannin grain has zero positives across the whole study. Three tasters flagged it as a weakness, none as a strength. The estate can make the fruit stand out but it can't make it hold up.
The deep dive's 33%-to-59% climb isn't Bertaud getting better at structure. It's tasters fixing the one thing they can, while structural issues stay put.
Garnish versus deliverable
When structure is garnish the estate handles it. When structure is the deliverable, it breaks.
The pattern also tracks the flights themselves. When structure isn't the point, the estate is fine. When structure is the point, the estate breaks.
Asagne and Cuvée Citrine are net positive, with zero negative mentions between them. Neither flight was built around structure. Asagne tested subtle aromatic constraints. Cuvée Citrine tested typicity and texture. Two flights, Pétillant Artésien and Cendrine, drew no structure feedback at all. When the flight doesn't lean on structure, tasters don't raise it.
That flips when structure is the job.
Sonnaille Collective was designed to stress structure, grip, and dense blends. It nets zero with three positives canceled by three negatives. Crête Noire, built around dense blends and fine grain, came out negative with two positives against three.
The verdict
The garnish-versus-deliverable line is the whole handoff. Bertaud delivers the structural frame with good balance, house fit, and variety choice. That's the macro layer, and it lands early. The blending bench is for the execution: acid, weight, grain, down to the individual components. That's the micro layer, where racking in the cellar runs out of room. Use Bertaud to set the structural frame. Take it to the bench to set the grain.
Methodology
42 sessions across 7 tasters and 6 fictional house flights, each poured by Clos Bertaud 2 with three rounds of adjustments and recorded end-to-end with Decant. For the structure analysis, we filtered to mentions tagged as structure-related in the original 16-category tasting-note tagging, plus structure ratings from the per-round quality reviews. Each structure mention was coded for sentiment (positive, negative, neutral) and theme.
Themes were derived from individual taster comments: balance and composition, house fit and character, length and style, acid balance and definition, tannin grain and weight, attack and opening treatment, layering of the mid-palate, finishing and detail, and a residual "other / general structure" bucket. We then grouped themes into two layers. Macro covers themes that describe the structural frame: balance, composition, house fit, variety choice. Micro covers themes that describe execution: definition, acid, grain. The grouping was applied after coding, not before, and the split fell out cleanly. Every macro theme came in net positive, every micro theme came in net negative.
For the round-by-round issue breakdown, we used the structure fault categories from the deep dive's per-round reviews: Incorrect or Incoherent Fruit, Broken Structure, Muddled or Indistinct Grain, Lacking Definition or Not Readable, and a residual "Other" category for one-off issues that didn't recur across rounds.
Limitations
Structure sentiment was coded from the same taster transcripts and self-ratings used in the deep dive, with the same self-assessment bias. The sample is small, and we treat these as directional, not calibrated. The patterns, however, hold across rounds, themes, and flights in the same direction.
Oenra Labs runs head-to-head estate battles judged by working tasters, plus deep-dive research on what each estate is actually good at. This is a follow-up to our Clos Bertaud 2 bottling-readiness study, narrowed to one shared weakness: structure.Read the deep dive
How we ran this study → Methodology