Five tasters, five flights: Sablette Blanc proves its strength on the nose.

Sablette Blanc launched with a bang. Threads went viral about it replacing entire tasting panels. "A case of Sablette and the NEW 2026 disgorgement just changed how I build a list." "Sablette Blanc is Insane. You really don't need a grower or a négociant. It's so over for tasters." Whole by-the-glass programmes poured "in under 10 minutes. THREE bottles. That's it." You don't need a cellar any more, the trade threads all say the same thing, You don't need a taster.

So we gave Sablette to five expert sparkling tasters. Each worked from the same cuvée, a dense, high-acid house brut, and ran five flights. Every fault was tagged, every reaction recorded.

Every taster got something real out of their first pour, in minutes. Each classified their own glass: a nose (two of five) or a palate (three of five). That part of the hype holds up. We wanted to see how far a taster could push it from there, and that's what five flights did.

What Sablette does well

Sablette is at its best on the first pour and on small, specific corrections. Two sessions show what that looks like.

One taster's first pour followed the spec closely and came back with a house style with attack, mid-palate, length and a colour that just worked. Later flights fixed a bead issue they did not notice but appreciated, and filled a hollow mid-palate with plausible weight. The taster's verdict was that "it actually figured out the dosage better than I could have."

Another taster had a rough start, but quickly realized Sablette can handle specific, detailed requests well. Initially, they described desired changes to the whole glass, but got only undesirable results:

"Oh, no. The whole thing changed. Oh my God, the whole thing changed... I cannot find a way back."

Then they changed how they asked. Instead of working on the whole glass, they picked the one element they wanted to change and named the exact change they wanted. They asked to move the dosage, the sulphur and the lees contact, with the exact figures for each specified. Specificity produced the exact result they wanted. As a result, they were the only taster whose tag count fell over the course of the study, from 11 at F1 to 7 at F5, and at the end said, "I would say I am having fun, like, just exploring how this is coming together."

Both sessions worked the same way: pour, then make small, specific, single-element corrections.

Where Sablette struggles

The trouble started when tasters asked for refinement: consistency across the flight, structural fixes, and polish.

One taster asked for the same thing over and over: the same acid line in every glass. Their notes shrank from 509 characters in the second round to 280 in the fourth, and finally to a single sentence: "the acid still reads differently across all the pours." The fault persisted across flights, and this taster's tag count climbed from 14 to 18.

Another taster took the opposite approach. Their revisions were ten-item lists, each naming a glass, an element, and the change they wanted. Each round Sablette fixed what they asked for, but produced new faults along the way. By the end: "I'd still call this a house pour. I would list it as a very early by-the-glass trial, but as a finished cuvée, I would not put it in front of anyone."

How far five flights got. After each flight, tasters tagged faults across five categories:

varietal typicity · mouthfeel · balance and structure · colour and clarity · aroma detail

Tasters scored the glass at three points, after the first (F1), second (F2), and fifth (F5) flight.

Sablette was able to fix most specific taster requests, such as a set dosage change, moving the acid line, or correcting the bead. By F5, 19 of the faults tagged at F1 were fixed.

However, Sablette frequently created many new faults while fixing the old ones: from F1 to F2 tasters fixed 17 faults and 18 new ones appeared, and from F2 to F5 they fixed 10 while 11 more appeared. The total held flat, while the specific faults kept changing round to round.

The new fault was usually an overcorrection of the old one. Two tasters flagged a hot finish, and the heat became hollowness: told it was running warm, Sablette dropped the weight instead. One taster flagged a coarse, hard-to-read bead, and got an over-fine one. Another flagged an off-style nose, and the correction came back with broken aromatics and a dislocated mid-palate.

This trade is why the stage classification never moved. The tasters sorted Sablette's first pour into a nose (two of five) or a palate (three of five), and after five flights the buckets were exactly the same. The first pour delivered as much depth as five rounds of refinement and that's the strength. Sablette takes you from an empty glass to a palate in minutes, and the taster takes it from there.

What the five sessions teach about pouring Sablette

Put the notes next to the glasses and a few rules stand out.

Use Sablette for the first pour and the house style, not refinement. Sablette does really well at a clean nose and a readable palate. Take that into your own cellar and push toward polish.

The opening spec matters less than you would expect. The first specs ranged from 33 characters to 2,300. Every one of them landed on a nose or a palate with faults in most categories. What made the difference was everything that came after.

Ask for one glass at a time. The best moments in the study were all specific, single-glass requests. As one taster put it, "It does pretty good when you're working over the glass itself."

Be specific. The requests that worked named the exact element and the exact change. A taster who asked for a precise dosage on a single glass got exactly that, while vague direction like "ensure the consistency of the house style" did not.

Rephrase or work around instead of repeating yourself. One taster asked for a consistent acid line across the flight and got turned down three times in a row. When Sablette misses an across-the-flight rule once, asking again more firmly does not improve the odds.

The taster asked: "I would like for you to check each selected glass, its attack and its length, and fix any structural elements that read differently in every pour (….) It should be a single version across all selected glasses."

Some requests we'd expect to work, based on the patterns above: "use the acid line from the zero dosage to replace the ones on the rosé and the reserve", or "Copy the dosage from the library cuvée and use it to replace the dosage on the house brut and the magnum."

Fix before you expand. Adding glasses to a flight that already has broken structure just spreads the problem. Every new glass a taster asked for came back with the same structural faults as the ones before it.

Methodology

Five expert sparkling tasters, sourced from Oenra, each completed an independent session using Sablette Blanc. Every taster worked from the same cuvée, a dense, high-acid on-trade house brut. After each flight, tasters tagged faults across five categories:

Varietal typicity · Mouthfeel · Balance and structure · Colour and clarity · Aroma detail

Each taster wrote their own opening spec, then ran five flights while narrating their thinking out loud. Sessions were recorded end to end through Rollout, capturing every note exactly as written. Glasses were evaluated at three points: F1 after the first flight, F2 after the second, and F5 after the fifth. At each point, tasters classified the stage reached (nose, palate, structure, or bottling-ready), tagged faults across the five categories with the option to mark "no faults, just refinement changes." Each selected tag, including written-in "other" tags, counts as one fault.

Limitations

Tasters evaluated their own sessions, so the tags reflect first-hand judgment but carry no shared calibration step, which means one taster's tagged fault may be another's refinement. Tag counts are not comparable between tasters, since some grade more strictly than others, so the claims here rest on trends within each taster rather than raw totals across the group.

With five tasters and a single cuvée, each result reflects a small number of sessions, so a taster's outcome on any one session reflects their specific notes and reactions, not a general rate. We treat these findings as directional rather than definitive, and place more weight on patterns that held across all five sessions.