Very gratifying to see the interest in the first entry. Truth is, there are many unknowns with this style. One interesting proposed idea, was that the item was sewn with the CB and shoulder reversed. I can see how that could happen based on the piece’s shaping but even if it were true, the underlying problem -that of the CB neckline length being too short- would not have had such a dramatic effect.
Some also mentioned the issue of a right and left back yoke (or whatever we’re calling the shoulder thingy), with the implication these were separate pieces. This piece is a two per, the same on each side. The only time you have a right and left one per for a button close is if there is some sort of unusual button closure acting as a design feature. That’s not the case here, it’s a functional closure.
Some mentioned the problem lay primarily with the left shoulder due to the larger gap between the yoke and the shoulder tip. Maybe, maybe not. We can’t know. I suspect not (I would eliminate the most obvious first) and for several reasons. No one mentioned that the degree of slope of each shoulder differed. This is true of all humans. That the right shoulder is higher means the child is right handed or she may be standing with her weight on her left foot. It is still pulling on the right shoulder, just not as dramatically as on the left. More importantly, deciding the scope of appropriate correction matters depending on whether you’re making a custom item for one person -then you would adjust the shoulder of each side differently- or if you’re making RTW (as most of us do), you don’t have this luxury. Sure, you could do it since the majority of consumers are right handed but someone left handed could not wear this dress at all. Being left handed means having a higher left shoulder and if the left yoke is cut at a more extreme angle than the right, it would be impossible to wear.
Several people mentioned making two corrections at once. You’re free to disagree but I avoid doing that at all costs if the two have the same underlying cause or there is enough of a correlation between the two that one could reasonably suspect they are related or are one and the same (it is a narrow span piece). The problem here is that the back neckline edge is too small, the line needs to be longer to travel the distance. That seam line is so small it is causing the shoulder edge to flare as the garment seeks the path of least resistance. You can’t know how too-small something is if it is too small. It is impossible. You can only adjust what is too large. So, you must first make the too-small area larger in order to know how to correct it and then if the other area (the shoulder tip) is too large, you can then pull that in. I know it saves time making a two step correction but I’ve learned the hard way that it often creates more work later (a further correction is needed) rather than less, again, if you suspect the cause of each needed correction shares a root cause (as I believe this one does).
In summary, this is not to suggest that others suggestions would not have bearing as a later course correction but the first step is to open that shoulder seam closest to the neckline to gauge the spread (how too small the back neckline is). One could pin a strip of muslin underneath, pinning the dress shoulder seam to it until the button stand relaxed. Once a problem has been properly defined, the simplest solution is usually the correct one. Do that first and then worry about anything else that remains amiss.