The issues with that was already outlined in my post, but I will repeat it her for ease of access.
First step: Identify the states that need to be flipped and then organizing the ground game
Problem: It is hard to know, beyond polling (which is already error-prone), which states need flipped. The issue with Ohio is that while it is a swing state, it generally isn't the Tipping Point state. The tipping point state is defined as the state, won by smallest margin, that was necessary in getting to 270 electoral votes. In 1992 the tipping point for Bill Clinton was Tennessee, which took him from 263 ECV to 274 ECV. He went on to win 10 unnecessary states totalling 370 ECV. In 1996 the tipping state was Pennsylvania, pushing Clinton from 256 ECV to 279 ECV, he then won 9 unnecessary states giving him a grand total of 379 ECV. 2000 of course being everyone's favorite example had Florida as Bush's tipping point state (he won no extra states). 2004 had Ohio as the tipping point state, pushing Bush from 254 to 274, he then won 2 extra states. 2008 it was Colorado, and Obama one 5 extra states, 2012 was Colorado again with Obama only winning 3 extra states.
So in 1992, you'd need to not only know the 11 states that went for Clinton by the smallest margins, you'd also need to organize around 750k fradulent votes.
1996 was 10 states and 1.4 million votes
2000 - again weird year - was 1 state and 500 votes
2004 was 3 states and 134k votes
2008 was 6 states and 1.1m votes
2012 was 4 states and 527k votes
2016 is shaping up to be 7 states and 1.2m votes (based on current polls)
A further issue is any state with safeguards against voter fraud, you may either need to target harder states (increasing risk as you need more and more votes to flip those states), or produce false credentials to circumvent the security (again increasing risk of being caught)
Also these are the bare minimum, if you fraud those exact number of votes, every state that you organize fraud in will trigger automatic recounts, so you need to increase those voting numbers by around 2% to get around auto-recount.
Second Step: Setting up the ground game
Problem: So you need to target specific precincts, keep turnout numbers at the same level, and leak small amounts of votes into each precinct. If you flood a few precincts, you risk flagging high voter turnout and then triggering an investigation. So your ground game needs to strategically target specific precincts, susceptible to voter fraud, and spread out your vote.
You also need transportation for your duplicate voters. You need to ensure they'll actually vote multiple times and not just go along and not vote. You need to ensure if they do vote multiple times, they are not voting back and forth (i.e. 5x Clinton and 5x Trump so no net gain). You also need all of these people to stay quiet, and not break under pressure as everyone questions how polls could be so wrong in so many states.
If you hire a middle man to organize all of this, you need to make sure he's doing his job, and it'll be expensive. You also need to keep this information off your tax records and do it all under the table. Your middle man might just pocket the money and say he's doing it. You can't actually check without getting implicated. If your guy gets caught, there's a good chance he'll rat you out if the FBI gives him a decent deal. You need to account for all of that.
Third Step: Voting Day Execution
Problem: Everything needs to work right. There is no second attempt. Anything goes wrong, anyone leaks this scheme, around 5% of votes (in any state) don't go through due to various problems like time, transportation, people backing out etc, and your plan fails. So much can go wrong. It's why organized crime is so rare and easy to topple.