You can't get a super-accurate measurement without an integrating sphere, but if you want to invest a small amount of time in it, you can get much more accurate than a ceiling bounce test by making your own light box. You can search here on CPF to find some threads on how other people have made their light boxes, or if you light I can give you details on how I made mine.
Once you have a light box made, you can sort of "calibrate" it. First you test various lights in it that you can find reliable measurements of the lumen output from other people's integrating spheres, then you can compare the measurements of the real lumen output to the readings from your light box, and find a conversion equation. The more lights that you have that you can get reliable measurements of real lumen output, the better your equation will be, and the better you'll be able to estimate the output of the lights that haven't been measured in integrating spheres.
If (like me) you don't already own a ton of lights that other people have measured, you can still use a light box for more precise measurements (than a ceiling bounce) to compare the overal output of the lights you do have, you just won't be able to convert that to lumens.