Economists mostly expect the Fed to tighten in September but market expectations flip-flop between September and December, shifting odds based on each piece of data or Fed comment. For instance, the market gave slight credence to this week's weaker-than-expected ADP report, showing payroll growth of 185,000, but gave high credibility to a jump in the employment component of the ISM Nonmanufacturing survey to a 10-year high. LaVorgna said that, even though weaker, ADP has been trailing the government payrolls number and is in line with a more than 200,000 increase in jobs.
The July monthly payroll data will be released at 8:30 a.m. ET Friday, and economists forecast an unchanged unemployment rate of 5.3 percent and average hourly wages growth of 0.2 percent, according to Thomson Reuters. That compares to June's job growth of the same 223,000 and flat wage growth.
"If this number is strong, it's going to have a larger reaction than previous numbers because they're running out of time," said John Briggs, head of strategy at RBS. Briggs said if the jobs report is better than expected, the 2-year note yield could quickly move to the high end of its range or above. It was at 0.7 percent Thursday.
One piece of jobs-related data that has shown positive consistency is weekly jobless claims. The latest report showed 270,000 claims, a slight increase but the 22nd week that claims held below 300,000. During the survey week for the July employment report, jobless claims reached 255,000, a 42-year low.
Read More US weekly jobless claims total 270,000 vs 273,000 estimate
What it will take to get the Fed to move is a sense of consistency: The economy does not need to improve, but it can't get derailed. Economists say the Fed will also weigh key components of the July employment report this Friday and the August report on Sept. 7, in addition to wanting to see nonfarm payrolls of 200,000 or more.
"The unemployment rate is probably the more important one. Month to month, it does bounce around," said J.P. Morgan economist Daniel Silver. "The Fed has said that's one of the better measures of slack and it's getting close to where it should be at least at the end of this." Fed officials forecast the unemployment rate at 5.2 to 5.3 percent at the end of this year, and near 5 percent is viewed as full employment by many Fed officials.
Another important measure for the Fed is wages but the Fed has indicated it could lift rates even before wage growth picks up. "The labor market is getting better, and they think that's going to continue. I think they're going to move without seeing a big pickup in earnings," said Silver.
But Mark Zandi, chief economist at Moody's Analytics, said if average hourly wage growth doesn't pick up to about 0.3 percent this month, the Fed may not see the firepower for a rate hike. "I'm expecting a strong gain, 0.3 percent, maybe even 0.4," he said, noting that the flat reading last month may have had to do with timing. "My forecast is 0.3 percent (wage growth) and 215,000 (nonfarm payrolls). I'm expecting a rebound. If it's 0.2 and 185,000, I don't think that's sufficient. We'd have to see what the August number looks like," he said.
Read MoreWhy oil prices could stay lower for longer
Market expectations were high for a sign of rising wages in last week's second-quarter employment cost index. But Zandi said the disappointing loss of momentum in the second-quarter ECI was not meaningful, even though the futures market immediately gave less credence to a September Fed move based on it.
The index showed a gain of 0.2 percent after a surprise jump of 0.7 percent in the first quarter. "It was a fluke…. It was all concentrated in the Northeast in services, and it comes after a very strong first quarter," he said. When accounting for the first quarter's outsize increase, "it continues to show slow but steady acceleration."
August's data, however, may pose its own problem. LaVorgna points out that August nonfarm payrolls in each of the last four years negatively surprised the markets by an average 55,000 jobs. Each report was below 200,000. He points out that while he continues to expect a September hike, it would be worrisome if there were a weak report Sept. 7, just before the September Fed meeting.
Read MoreWill jobs report tip the Fed's hand?
"History suggests that even in years when job gains are sturdy, we tend to get at least three soft reports," he noted, adding they typically have come in below 200,000 before revisions. March's report was the weakest of the year so far and was initially reported at 126,000.
Zandi said the seasonal factors in August include companies reporting data late because of summer holidays. "The August report is oftentimes bizarre. It was back in 2011, we got a zero print. That is something watch for, for sure," he said.