You better post a link to that solution.

A common red LED may have reasonably known wavelength and intensity. But that would not be a "red LED" but a very specific LED driven at a very specific current. And not only that - the variance between different batches of diodes will vary a lot. Not only will the performance of the silicon wafer vary. But the manufacturing process will may also introduce mechanical tolerance problems in the direction the little chip is oriented, which will affect the amount of light the LED sends out in the main axis direction.

But that does not solve the problem of measuring color. It just helps to measure the intensity of that specific wavelength range the diode sent out, in a situation where there are no other wavelenghts represented. And the next thing is that a red LED does not has as pure color as a red laser, in which case the LDR sensitivity to different parts of the red spectrum may introduce significant measurement errors.

In real life, you would probably measure the full spectrum in 100 narrow but overlapping wavelength bands. By being overlapping, the pure light from some light sources will not be missed. By being narrow, you will also make sure that a red spectrum band will not falsely detect green light. By having many narrow bands, it is possible to take absolute energy readings and convert them into a perceived after compensating for the sensitivity of our eyes.

For a low-end solution, it would probably be easier for you to use a web camera or similar and analyze the image. But that would be a low-end DIY solution. The real solution would be to use a sensor intended for the task. And only if the gual is to produce and sell a product in many-k volumes.