Dear Slack,
I write this article as an intervention to a loved one.
Slack, I love your product, but your search is in desperate need of rescue.
It would be one thing if you didn’t tout your searchability. But Slack itself stands for
“Searchable Log of All Conversation and Knowledge.” You yourself recognize that search is paramount to your identity.
As an avid user of your interface, I know that a Slack search often yields results that feel like a ’90s web portal—words are matched, yes, but the order, and therefore the very important aspect of correct context, seems unimportant. Synonyms are completely unsupported. User intent is ignored.
Like many others in the tech space, my company uses your services for our internal chat. We came on board in 2015, enticed by your organizational features, seemingly limitless capacity for conversation history, and app connectivity.
Importantly, we have no desire to make a switch. We would even love it if you could replace email entirely. But if a messaging app with a better search function emerges, we will have to leave you.
And we won’t be the only ones.
The average worker spends 5% of their total screen time on messaging apps like Slack; and according to the International Data Corporation, the average person spends two and a half hours a day searching for information. If someone were to enter the market—perhaps someone who owns a search engine and has the expertise to decrease that soul-crushing search time—why wouldn’t we switch?
There’s a lesson that tech companies have repeatedly learned in the past 25 years: If your search function is inferior, your users will eventually abandon you for a competitor with better search. When Google appeared on the scene in the late ’90s, Yahoo, Excite, Lycos, AltaVista, and a multitude of other web portals came up against this same roadblock.
Each of these sites had its own search functionality, along with a multitude of other features and curated content, which Google lacked. But Google knew what was most important to users: the best search. Now of course it’s the world’s dominant search engine, while those other companies are but a footnote in the history of search engines.
Good search doesn’t just make good business sense. Google’s accurate results are so ingrained in our society that those same history books will, as the Guardian’s John Naughton puts it, “see our era divided into two ages: BG and AG—before and after Google.”
A 2009 (AG) study found that using a search engine excites regions of the brain that handle decision making, vision and complex reasoning—and those effects are nearly doubled among more experienced internet users. Dr. Gary Small, who led the study, concluded that the more we search, the more our brains react to searching. Google is literally re-shaping the way we think.
In the same vein, a seminal study from 2011 found that having access to good, trusted search allows us to hold less information in our minds that we know can be easily searched for. The authors even coined a term for this—”the Google effect.” In short, giving users better search means they can not only save time, but also save memory for more important topics that can’t simply be searched for.
Slack, we realize improving your search won’t be a small undertaking. But an investment in your search team is an investment in your users, both present and future. Amazon employs more nearly 500 people to work on search alone—far more than any of its competitors. It also captured 49% of US e-commerce in 2018 to the tune of $233 billion.
Yes, it was expensive to build search as good as Amazon’s, but clearly its investments have paid off. Two-thirds of Americans start their search for new products on Amazon.
I’m not sure how many people work in the search department at Slack, but I don’t think it is enough.
Maybe your search function yields antiquated results because it relies heavily on antiquated algorithms. Not to get too technical, but Slack runs largely on TF-IDF, or term frequency-inverse document frequency, a kind of prototype artificial intelligence which was invented in 1957—long before the birth of chat software.
Meanwhile, the likes of Google and Amazon rely on modern machine-learning search methods built on models that use users’ behavior after the search to improve results and eventually understand the intent of a user’s query. And that’s what the search experience should be like in 2019.
After your initial public offering in June, Slack, you’ve got more on the line than ever before. Though prices ballooned on the first day of trading, that momentum was short-lived. An improved search function is a sure way to excite investors—and paid customers, both current and potential.
That last group is especially important. Yes, Slack has at least 10 million active daily users, which is an impressive number to be sure. But of your 585,000 accounts, just 88,000 pay for your service, 575 of whom made up 40% of your revenue last year.
You have a lot of eggs in very few users’ baskets—eggs that will break if those users switch to a competitor like Microsoft Teams, which has 13 million active daily users, recently introduced new collaboration tools, and is rapidly improving its search.
Slack, I love your product, and I want to keep using it for a long time to come. Please fix your search so that I can.