Well I'm going to ramble a bit here, so be warned...
Speakers that have a great deal of variation in the shape of their radiation pattern (due to beaming) are less likely to give you decent soundstaging and decent tonal balance from well off axis.
Let's look at how the ears work a bit, just as background.
The ears localize sound sources primarily by both intensity and arrival time. Intensity is the primary mechanism above 1.5 kHz, and arrival time is the primary mechanism below 1.5 kHz, but both play a role in both frequency ranges.
Note that strong, distinct early reflections are especially detrimental to imaging. It's a good idea to either diffuse or (if necessary) absorb first reflection energy.
Now to get good soundstaging well off-axis is tricky. You nearer speaker's output naturally arrives first, and its output is also usually not only louder, but is in particular louder in the treble region. These factors combine to pull the image way over to the closer speaker even if you're just a little bit off-axis.
One way to minimize this effect is to use line-source-approximating loudspeakers, because sound pressure level falls off more slowly with distance from a line source. So for the off-axis listener, there is less loudness differential between the two speakers, giving a better soundstage.
Another way to get a good off-axis soundstaging is to toe the speakers in severely, so that their axes criss-cross actually in front of the listening position. This technique works well with speakers that have a slightly rising response on-axis. Now what happens is, the person off to one side hears the nearer speaker first, but because he's more on-axis of the farther speaker its output is actually louder (especially in the treble region). An added benefit of this approach is that, if the speakers radiation pattern is well controlled (by a horn or waveguide for instance), the first sidewall reflection may well be eliminated by simple geometry. The net effect is a more realistic soundstage from well off-axis. For best results, the speaker should have its flattest response not on-axis, but rather something like 15-25 degrees off axis, as this is where the listener will be sitting when the speakers are toed way in, so this is his first-arrival sound.
Yet another technique that works reasonably well is to use a speaker whose radiation pattern is approximately omnidirectional. This will at least give the off-axis listener a better presentation than beamy speakers pointed straight ahead or only toed in a little.
Speakers designs that produce a well-controlled, uniform radiation pattern tend to be the best for off-axis listening, provided those patterns are aimed properly (if the speaker is fairly directional).
Let's also look at tonal balance for a moment. Even if you can't get great soundstaging from well off-axis, the tone of the voices and instruments should still sound natural. This calls for a fairly wide, very uniform radiation pattern over as much of the spectrum as possible (the treble region is most critical here, and very few speakers have a uniform radiation pattern through the treble region).
I think it can be argued that there are really two "sweet spots" - the imaging/soundstaging sweet spot which will be best up and down the centerline but can extend off to the side fairly far with suitable loudspeakers and set-up; and the "tonal balance" sweet spot which can literally be throughout the room if the speaker has a suitably uniform radiation pattern.
The Gallos mentioned above have a wide imaging sweet spot and a wide tonal balance sweet spot due to their wide radiation pattern. The Martin Logans have a fairly wide imaging sweet spot due to their line-source, fairly uniform radiaton over a 30-degree arc, but their tonal sweet spot really doesn't fill the room as well because their radiation pattern is very different (approximately omnidirectional) in the lower octaves where they use a conventional woofer.
Personally, I place a high priority on a wide sweet spot, and perhaps oddly I place a lower priority on superb imaging right in the middle of that sweet spot. OFten there's a trade-off relationship between superb imaging in a small area versus decent soundstaging over a larger area. I have some familiarity with electrostats - the flat-panel InnerSounds have the best imaging but the smallest sweet spot; the 30-degree-arc Martin Logans not quite as good but the sweet spot is larger; and the 90-degree-arc Sound Labs (which I sell) don't image quite as precisely as the Martins but have a very wide sweet spot.
In more conventional speakers, you might want to look at designs that use horns or waveguides of some type (concentrics often work pretty well, as the woofer cone approximates a waveguide for the concentric tweeter), and at designs that approximate omnidirectional radiation.
If I had to guess based on your description of struggling to find the sweet spot, I'd say that you may have a problem with early reflections off a wall or piece of furniture, or diffraction off of a nearby edge (perhaps a bookcase or some other vertical feature). I think something is screwing up the early-arrival sound in your room and that's why you're having a hard time finding the sweet spot. Much as I think "new speaker fever" is a fun thing to have, try draping a towel or blanket over possible problem areas before you empty out your wallet.
Best of luck to you!