The Gordon Growth Model is a simplified version of one of several models examined by MJ Gordon in his 1959 paper. It values a security using the discounted value of future dividends assuming a fixed growth rate:
d/(r - g)
Where d is the next dividend,
r is the required rate of return, and,
g is the rate of dividend growth.
This is simply the usual DCF formula simplified for a fixed discount rate and a fixed growth rate. This tells you what is wrong with the model: why should these, especially the growth rate be fixed?
The implication of this is that the Gordon growth model is not at all suited to valuing growth companies as high growth will slow at some point. It is best suited to companies where the primary driver of future growth is GDP growth. This does not mean that dividend growth will equal long term GDP growth, just that it is ultimately driven by growth in the wider economy. Long term growth lower than (or equal to) GDP growth is more plausible than higher growth.
The required rate of return is the cost of capital for the security being valued: so if you are valuing shares it is the cost of equity capital.
Like other simple formulae the Gordon model is best suited to situations where the use of a fuller model is impractical: filtering to create a list of potential investments for further analysis, or for a pure mechanical investing strategy. Where possible, it is far preferable to use dividend discount model that does not makes such a drastic simplifying assumption.