Shares of Netflix (NFLX 4.17%) have done remarkably well this year, having climbed 53% so far in 2020. Roku's (ROKU 5.41%) stock has done even better, more than doubling Netflix's performance and gaining 129%. Going forward, however, their paths will diverge even further, with Roku remaining a buy and Netflix underperforming.

That's according to Needham analyst Laura Martin. On Wednesday, Martin named the two stocks Needham's "top pairs trade" for 2021, going short on Netflix and long on Roku, while retaining her $315 price target on Roku's stock.

A man pointing a remote at a TV.

Image source: Getty Images.

Martin cited several catalysts that she expects will drive Roku shares higher. The acceleration of cord-cutting during the pandemic resulted in a greater number of viewers for the platform's ad-driven content, at 46 million, up 43% year over year. The trend also drove advertisers to platforms like Roku as a way to respond to these shifting viewing patterns. Perhaps most importantly, Roku attracts the younger demographic of viewers most sought after by advertisers. 

For Netflix, however, Martin sees a very different future. She believes that many of the subscribers who signed up this year were merely pulled forward, stunting future growth. Additionally, when live sports and music return, Netflix won't be as appealing to some viewers. Finally, Martin believes that lower-priced competitors that offer live broadcasts -- like the upcoming Olympic Games -- will benefit at Netflix's expense. 

While I believe that Netflix's growth will slow from here, I certainly wouldn't recommend selling. Netflix is the streaming market-share leader with 195 million subscribers, remarkable pricing power, and consistent growth in international markets. 

I have no argument concerning Roku's future potential. In the third quarter, Roku reported revenue that grew 73% year over year, driven primarily by platform revenue that surged 78%, from solid gains in digital advertising, the Roku Channel, and licensing of the Roku operating system for smart TVs. 

Given the available evidence, I think Martin's only half right.