When Palo Alto Networks (PANW 0.84%) delivers its fiscal fourth-quarter results on Aug. 31, it will have to show significant progress in a couple of key areas if it's going to change my bearish opinion. And based on its marginal stock performance this year, I'm hardly the only one.

For years, the Street seemed more than content to focus on Palo Alto's skyrocketing top-line growth while brushing aside its mounting losses. Last  quarter's 25% revenue growth to $431.8 million not only beat Palo Alto's own guidance, it handily surpassed pundit's expectations. So what's not to love? Simple: Despite its stellar sales growth, Palo Alto remains in the red, and what improvement it reported on the profitability front was marginal at best.

Drawing of part of the Earth with multiple geographic areas circled.

Image source: Palo Alto Networks.

What to like

The top line should continue to show healthy growth this quarter assuming CEO Mark McLaughlin's guidance proves correct. Palo Alto is expecting between $481 million to $491 million in sales, which would be a 20% to 23% year-over-year improvement.

Don't be surprised if Palo Alto reports revenue on the high end of its forecast range, if not above it. McLaughlin wouldn't be the first CEO to miraculously beat his own sales guidance on a consistent basis. Reporting a revenue "beat" helps take the focus off off other, less rosy, results.

Among its wins are ballooning subscription sales, which drive its recurring revenue. While total revenue climbed 25% last quarter, Palo Alto's subscription  sales soared nearly 46% to $267.6 million, more than offsetting its meager 1% increase in product sales.

McLaughlin has declared that increasing Palo Alto's recurring revenue foundation is a key priority, and last quarter was close to a home run in that regard. Another of his objectives is to pare overhead and increase efficiencies, particularly as they relate to sales costs -- an area where the company continues to fall short.

What not to like

Operating expenses once again climbed last quarter, and we will likely learn there was a further rise when fiscal Q4 results are announced. At $357.2 million, last quarter's operating costs were 18% higher than a year earlier. Cost of revenue, while smaller in total, jumped 30% to $123.7 million. Though Palo Alto's $0.67 per share loss was 8% better than last year's $0.73 a share loss, there is still a long road ahead before its profitable.

Palo Alto should take a page from FireEye's (MNDT) cost-cutting book. While FireEye's 6% gain in revenue  last quarter to $185.5 million got the headlines, it was the significant progress the company made on CEO Kevin Mandia's efficiency improvement efforts that took center stage.

FireEye pared its operating expenses 24% last quarter to $178.2 million, and reduced cost of revenue, though marginally. Another bright spot was its 26% reduction in sales overhead; cutting that cost is a particular focus for Palo Alto.

But at Palo Alto, last quarter's sales and marketing expenses jumped 16% to $226.9 million, equal to 53% of total revenue. To its credit (sort of) last quarter's spending on sales was a decrease on a percentage basis from a year before, when Palo Alto shelled out 57% of its $345.8 million in revenue on sales and marketing.

The reason I don't like Palo Alto stock "yet" is that -- as mind-boggling as it may sound -- the 16% rise in costs last quarter was actually better than earlier periods. That may not seem like a lot to hold onto, but in McLaughlin's defense, the more intensive expense management efforts at the company are still relatively new.

What took Palo Alto so long to recognize its spending was out of control is anyone's guess, but in this case, better late than never. Though analysts will likely focus on the revenue growth numbers when Palo Alto reports at the end of this month, it is the company's spending that will determine whether or not it should win back long-term investors.