It’s Go Time (Premium)

Microsoft’s goal for Surface Go is as understandable as it is necessary: To dramatically expand the customer base for this premium line of PCs. But for Surface Go to succeed, it has to deliver on some basics. Key among them, of course, is performance.

The good news?

Surface Go is the right idea. The Surface PC lineup is excellent today, and I can hold up recent generation products—Surface Laptop, Surface Book 2, and the 2017 Surface Pro—as examples of some of the very finest PCs I’ve ever used. These products succeed on a very visceral level, something we typically only associate with Apple products. I love them despite some curiously dated components.

But Surface PCs are too expensive. And for the PC market to stave off its inevitable decline, it’s not enough for Microsoft or any other PC maker to only serve small but profitable niche markets for gaming PCs and other premium PCs. This market needs the quality of Surface at price points normal humans can afford. It needs to reach the mainstream.

This concept isn’t limited to PCs, of course. We see quality levels and specific features that were once relegated to flagship smartphones make their way down-market over time too. Outside of personal technology, these trends impact other markets too. Car makers expand their product lineups, and often add new entry-level models as existing products evolve and get bigger and more expensive. You just can’t ignore the mass market.

The question for Surface Go, of course, is whether the idea—a lower-cost Surface Pro—has translated into reality. There are many ways you can save money when building something like Surface Pro, and Microsoft has tried a few of them. For example, it has used lower-cost and lower-performance Core m/Y-series chips in base models.

Surface Go is different. The Pentium Go processor in this PC is some order of magnitude less impressive than the Core i5 and Core i7 processors that PC makers expect. The device, too, is tiny, with a less than full-sized keyboard on its Type Covers. This seems to limit its usage to child, or to those with smaller hands.

But I think this compromise addresses the real world needs that Microsoft sees for the device: It’s not a PC that most would use all day, every day. It’s for occasional usage. Perhaps in a school setting where different students will interact with a single PC in short time increments. Or a PC that one might tuck in a bag and then pull out only when they need a display/keyboard that is better than their phone. Or at home, where you might turn to it from time-to-time to check on email, look up something online, or whatever.

Whether Surface Go will meet these needs is still unclear. At least to me.

As you may know, I’m traveling, so I wasn’t able to receive a Surface Go review unit in time for this week’s release. But I asked Brad if he would perform a simple performance test, which is something I do for every Windows-based PC I review: Use Handbrake to encode the 4K version of the Tears of Steel video to a high-quality 1080p format.

I have performed this test on dozens of PCs over the past two years. Modern PCs with quad-core 8th-generation Core i5/i7 CPUs will typically about an hour (plus or minus 5-10 minutes) to perform this conversion. PCs with previous-generation dual-core CPUs typically took about an hour and forty minutes.

On the extreme ends of the test results, I’ve seen scores as good as 31 minutes, for an HP gaming PC, and my HP AIO completed this encoding in about 42 minutes. The worst score I’d ever seen was for a 2017 Spectre x2, which for some reason required two hours and 14 minutes to complete this conversion.

So what about Surface Go? How does a Pentium Gold processor with 8 GB of RAM and the faster SSD option perform on this test?

3 hours and 22 minutes.

I can hear some howling out there. That it’s not necessarily fair to test this delicate flower of a child’s PC in such a way, that the Surface Go isn’t designed for power-user tasks like video encoding, or playing modern 3D games.

I get that. But Surface Go is designed to compete with iPad. And I’m not sure if you’ve seen the kinds of games that are possible on an iPad today. But they are impressive, visually, and from a performance perspective.

Will Surface Go perform acceptably for a limited range of tasks that may include light word processing, browsing the web, checking email, and watching Netflix? Maybe: For now, I will need to rely on Brad and the few others who have actually used this PC out in the world.

But my worry here is that there is a community of Windows or Microsoft enthusiasts who will simply will themselves to believe that the central performance compromise of Surface Go is good enough. It’s something we see over and over again, as people evaluate lackluster products like Surface 3, Microsoft Edge (which, admittedly, is getting better), or Cortana, and then conclude that it works just fine for them. It’s good enough.

I’d like to see people make better decisions for themselves. To not just choose a product or service because it is made by a certain company. And, conversely to not not choose a product or service because it is made by some other company.

Put simply, it’s important to really know what you’re getting into here. A reasonably configured Surface Go with Type Cover will set you back $650, and you’re not going to find enough quarters in the couch to cover that bill. It’s inexpensive in the world of Surface PCs and other premium PCs. But it’s still a lot of money.

Choose wisely, my friends.