Ten "colony forming units". That's essentially ten single bacterium per 100mL of water, or one in ever 10mL. Essentially nothing.
It has to do with how you measure microbes: the actual "True" number in bulk is knowable only to God; instead you would have to squirt a series of dilutions onto a plate (petri dish) with food that allows a colony to grow, that you can see with the naked eye. Then work backward through your dilutions to estimate an actual concentration.
The specific assay will be different when starting with a dilute sample and probably on a microwell plate (96 tiny petri dishes in one). For a 100mL sample, you'd need ten microwell plates, for 1000 x 100uL "wells", and 0 - 10 wells that grow bacteria. Plus controls. At that level, you don't know if it spilled off someone's face or not.
Or just spin down 100mL of sample, miniprep, and quantitative PCR. I don't know what's cheaper.
Then again a guy could take a swipe of every 100,000 nutrient bars or whatever, and grow on selective media, and compare to some old standard.
Bottom line is 10cfu is probably the "limit of detection" or "limit of relevance" for whatever assay they use.
Context is important: if it's in your tap water it means someone's shitting in it, or there's a leak from a septic system. It could be used as a measure of how well a municipal water supply is treating the water.
Keep in mind that E. coli is necessary in your guts and usually harmless outside your guts. A very few strains are pathogenic.
E. coli exists by the trillions and trillions and trillions in your guts right now, and you probably launch more than the population of humans on Earth into the air every time you flush your toilet.