In early 2015, processor giant Intel (INTC -2.40%) reported that its Mobile and Communications Group had posted an operating loss of $4.2 billion. This followed a loss of approximately $3.15 billion in 2013, thanks to the introduction of the company's "contra-revenue" program in a bid to quickly gain share in the tablet processor market.

With a reasonably sized footprint in the tablet market established during 2014, Intel has shifted its focus to trying to, over time, bring this business to profitability. The company told investors last year to expect an $800 million reduction in the mobile loss for 2015, and according to CFO Stacy Smith, the loss reduction for the year will likely wind up being closer to $1 billion.

Additionally, the head of the company's Client Computing Group, Kirk Skaugen, said mobile losses in 2016 should come down by $800 million.

Let's take a closer look at what drove the reduction in 2015 and what management expects to be the key drivers of the reduction in 2016.

In 2015, the loss reduction was due mostly to investment reductions
According to Smith, the reduction in losses during 2015 was attributable to two key factors: product margin improvements and "lower investment" (in other words, cuts in research and development, marketing, and so forth).

The product margin improvement was responsible for about a third of the reduction. Such improvements include the tapering off of the "contra-revenue" program that began in 2014 as the company transitioned its low-end tablet offerings to the more cost competitive SoFIA parts, which integrate both an applications processor as well as a cellular baseband.

The remaining two-thirds of the reduction, Smith indicated, came from a reduction in mobile-related investments. Intel has indicated that it will rely more on partners (with the two main ones being Rockchip and Spreadtrum) to try to build/distribute products based on its intellectual property.

In 2016, it's more about product margin improvements
Next year, there will be a sort of "role reversal" vis-a-vis the different "buckets" in which the loss reductions will come from. Smith says a third of the loss reduction next year will come from lowered investments while the remaining two thirds will come from product margin improvements.

Image source: Intel. 

As far as product margins go, Intel will likely benefit from a full year of shipping its much more cost-effective SoFIA products into the low-end of the market, rather than trying to ship Bay Trail chips along with contra-revenue subsidies.

In his presentation, Skaugen was optimistic that unit shipments of both SoFIA products as well as its stand-alone modem products would grow in 2016, which should certainly help the product margin story.

A preview of 2017
Although Intel didn't provide a forecast for 2017, I think investors should expect another significant loss reduction in the following year. The bad news is that Intel probably won't be able to cut costs much, if at all, from 2016 levels; at some point, if you cut too much in the way of research and development, as well as marketing expenses, it becomes nearly impossible to compete.

Any improvement that Intel sees, then, in 2017 will, in my view, need to come from product margin improvement. Fortunately, there are a number of things that should work in its favor there.

First of all, Intel is planning to bring its SoFIA products into its own factories on its 14-nanometer process by the end of 2016. This should enable both cost structure improvements relative to the SoFIA products that the company will ship in 2016 (since the current SoFIA products are built by third-party foundries that collect their own margins) as well as enhanced product competitiveness.

I think, though, that the "enhanced product competitiveness" is the more important factor. The better Intel's products are (and, by extension, the derivatives of those products designed and distributed by Intel's partners), the more units the company will be able to sell at reasonable prices/margins.

An increase in unit shipments thanks to more competitive products should also lead to increased revenue and ultimately margin dollars, driving meaningful profit improvements.

Additionally, although I am not all that optimistic about how Intel's XMM 7360 modem will fare in the marketplace, I do think there is a decent chance Intel can deliver a competitive product with the follow-on part known as XMM 7460.

In that case, I could see Intel winning a "second-source" deal to supply modems to Apple for its iPhone/iPad products that launch in the late 2017 timeframe. To support such a launch, Intel would need to begin shipping product to Apple by mid-2017 (if not sooner), which could potentially lead to much improved revenue/margin dollars in mobile for the year.