Paul Armstrong runs emerging technology advisory Here/Forth.
Despite a good second-quarter earnings call, Twitter is still labeled as a risk by financiers. People still aren’t clear about what Twitter wants to be when it grows up. CEO Dick Costello will not rule out an algorithmically led newsfeed that would adopt a Facebook-like approach to choosing only to display certain types of content.
Twitter has already had a redesign to emulate Facebook profiles, but adopting the same algorithmic approach to its stream is the wrong way to go. Here’s why Twitter should avoid this all costs:
Real-time serendipity is its USP.
If Twitter were to implement an algorithmic feed, it would lose its point of differentiation that would likely damage its de facto real-time information/news status unless a greater value proposition was offered (although it’s hard to see what this would be). Both as a professional tool for journalists and a point of record for regular users, Twitter offers a totally different kind of feed than say Facebook does because of its unfiltered stream of forced brevity. Every tweak with photos, cards and potentially new sell buttons drastically changes the delicate balance on the nuance — a nuance that should be protected.
Twitter is noisy, but it can be calmed.
Twitter’s newsfeed is an untouched chronological list of the world’s woes, wins and (sadly) monotonous wibblings — but noise is both the curse and the beauty of the service. Rather than resort to algorithms, why not make a concerted effort to simply educate its users? Lists, hashtags, advanced search and mute are all natural solutions to the problem. Put another way, when was the last time you saw a tip about advanced search and not a box coaxing you to “Invite your friends!”?
Twitter simply doesn’t know enough about us to make such decisions.
Twitter’s level of personally identifiable data is incredibly small compared to the big blue data whale that is Facebook. Granted, Twitter has the ability to scrape and analyze what we’ve said, when we’ve said it and who we’ve followed but that’s still big-bucket data. Even with the extra inputs (acquisitions, features) they continue to add, the amount of overtly given (or available) data is lower than the smallest amount of sign-up data points we see on Facebook due to the different privacy demands of the two networks. Thus, no algorithm Twitter could come up with would ever be likely to get the right mix of sense and insanity that the average user craves. Fifteen percent of people who “use” Twitter never have the opportunity to be served an ad anyway — an opportunity and an issue if more than 1:10 of your users isn’t adding to your success.
We already know algorithms can be bad news for brands.
If history teaches us anything about the flow of information, it’s that those who control the tap usually get paid. While I cast no aspersions on any social network out there, it’s hard to disagree that there is no financial incentive to control the amount or what users see when ad dollars are at stake.
As the amount of information being input increases, the ability to manage it is not only important for the future success of Twitter but a critical factor. Whether this means more information (e.g., Twitter Cards) or simply a quicker way of getting to a purer stream of information has yet to be seen — a combination of both with the right level of education is likely to help the fickle user base increase the much-coveted “time spent” metrics. However, without knowing what it is aiming for, Twitter’s future is unclear, and that will likely continue to worry investors and confuse both users and brands.