The Seven Deadly Sins are a good framework for thinking about your dream job. There’s the job you’d do out of greed — something involving either lots of options or the numbers two and twenty. There’s the sloth-job, probably manning some counter during the night shift of an institution that promises 24/7 availability but isn’t very popular after dark. Lust, Gluttony, Envy, Pride, all easy.
And then there’s wrath.
If I could take one job just to indulge my temper, I’d run Google Finance, because everything about it makes me so angry.
Imagine you’re describing roughly what Google Finance is, to someone who’s never used it but who knows what Google is. You’d say:
- It’s a site that naturally defaults to a search format. Bloomberg, FactSet, Money.net, Koyfin — every finance site presents you with a search bar and expects you to plug in the asset you want to search for. Google is pretty good at search.
- It’s a free site that monetizes with ads, run by the company that is the best in the world at monetizing free stuff through ads.
- It’s a site that aggregates information, from the company whose mission is to organize the world’s information.
- It’s a site for people with money, run by a company where entry-level engineers earn over $200k.
If you knew all that, you’d assume that Google Finance was the single best free site in the world for information about stocks.
You’d be wrong.
What Google Finance Gets Wrong
It’s a Google Product With Terrible Search
Let’s say it’s your first time on Google Finance. It’s a Google product, so you do a search. How about a large company with millions of customers, tens of thousands of employees, and $40bn+ annual revenue.
Granted, Google Finance is usually better than this, and they’ve improved a bit in the last few months. Still, it’s extraordinary. In the US, a ticker is a unique identifier — DAL is always Delta Air Lines.
Not to Google!
This shows up for a few other companies, too, generally companies whose tickers are also nouns (but not always! “EAT” doesn’t work, but “YUM” does. “FUN,” “SIX,” and “SEAS” give an industry trifecta of null results. “GOLD” and “AU” both fail, too.)
I put this one first because it goes beyond head-scratching and straight into gobsmacking territory.
If Larry and Sergey were dead, this would make them roll over in their graves. Google.com can match queries even if you egregiously misspell them.
(I wonder if the unit of measurement for comparing two typo-correction systems is literally BAC.)
What’s troublesome about this is that it speaks to serious ignorance about the finance-site workflow. Compulsive traders — the ones who will be on your site at all hours of the day — cycle through quotes nonstop. It’s basically their flow state. Unless you’re seriously talented, this is a bad way to spot opportunities, but it is kind of zen. Zip through a bunch of tickers, free-associating all the way: airlines are mostly up, so how’s oil doing? Oh, not so great, maybe we’re risk-off today; our friends TLT and GLD can confirm.
If you get this interaction wrong, you break the spell, which means you lose the user. Financial software providers invest immense effort in making it easy to zone out and check quotes, because that’s something their best users love to do.
In fairness to Google, most of the examples I gave offer a dropdown of autocomplete suggestions, and the suggestion usually includes the company you’re looking for. (Typing “HA” does not offer Hawaiian Holdings as an option, suggesting that Google is guessing matches based on company names, and ignoring tickers entirely.) But any time you take an interaction that’s always type-then-hit-enter and change it to type-then-maybe-scroll-or-click-enter, you lose something. Any UI that makes users think about the UI, rather than about what they’re trying to accomplish, is flawed.
Sometimes you encounter a weird design decision and you have to think about it for a while before you say “Ok, I know what you’re trying to do here.” But in this case you think about it for a while and say “Ok, you didn’t even try.”
Google Finance Doesn’t Learn Autocompletes
One thing that MOMA screenshot above proves is that Google is good at learning from typos. Really, really good. Spelling correction turns out to be an Alpha Nerd topic for Google. Their Alpha Nerd chief scientist, Peter Norvig, whipped up a simple spellchecker in Python just to show how easy it is. A few years later, they pulled off the notorious Bing Sting, when they noticed that Bing’s spelling corrections were more Mountain View-tier than Redmond-level.
The first time Google Finance failed to show me a quote when I plugged in a ticker, I thought back to the Bing story and assumed Google would fix it in a few weeks. After all, if their logs show that I searched for “DAL,” in Google Finance, then went to Google.com and searched for “DAL stock quote,” and then, just to twist the knife, clicked through to Yahoo Finance, surely they’d make the connection.
But they haven’t.
Here’s the crazy thing. Google is fastidious about logging everything. Once, in 2008, they put up a page that allowed people to search the index as it existed in 2001. So I’m sure they have logs of what people searched when they got bum results, where they clicked next, and when they closed the tab. An ambitious intern could code up a prototype solving 99% of these bum queries in an afternoon.
(The 90% solution: `if query in tickers_to_quote_urls: return tickers_to_quote_urls[ticker]`
Google Finance Ignores IPOs
No screenshot, sadly, but: next time a company goes public, try to pull up the stock price on Google Finance. There’s a very good chance it won’t be there.
This might be a subset of the autocomplete issue. Perhaps they have the quotes, but the newly-IPOed company doesn’t have enough quote-request history to populate the autocomplete roster. But it’s a flaw. For most companies, IPO day is the single day their stock will get quoted the most frequently; it’s the day CNBC will be repeating the ticker symbol over and over and over.
Old media is there. Google is not.
Google Finance Promotes Garbage News
If you went straight to finance.google.com yesterday, here’s what you saw:
Yesterday morning, that Elizabeth Holmes link led to a 404 error. Hours later, it’s still the top suggested link, and it still leads to a 404.
I can guess why the link is still up there. It’s very clickable. But a homepage should not be optimized for clickthrough rate; it should be optimized for something closer to expected session length, and expected session length over time.
What users are expecting from this page is, almost certainly, a summary of what’s going on in the market today, with maybe a few headlines about important stories. This is not hard content to find! Literally dozens of sites have someone on the “write what indices/rates/currencies/Apple/Tesla did today” beat. It’s not especially valuable content, so you don’t have to pay anyone very much to produce it.
Google Finance could, pretty trivially, find a way to categorize articles as market news summaries and place them front and center, which would prompt users to read the articles and then look up more information on the companies being talked about. This is the natural loop you want to start for low-intent visitors. If they show up and don’t (try to) look up a quote, give them stuff to look up. Instead, Google is aggregating clickbait from Forbes.com, which is itself mostly a clickbait aggregator whose sole purpose is to use the credibility of the Forbes.com domain name to game Google Search.
It’s one thing to fail, it’s another entirely to commit an own goal.
The content-recommendation own-goals continue. Because, like other finance sites, Google Finance aggregates breaking news about the companies it provides quotes on.
Facebook’s had some news flow recently. They paid a fine for misreporting video viewing data, which some people (incorrectly) believe tricked the entire online media sector into pivoting into video and then pirouetting into insolvency; they’ve had updates on a patent dispute with BlackBerry; they’ve had even more companies defect from their crypto-currency project; and of course they’re always launching new features.
What does Google Finance think the average FB investor wants to know?
Semi-generic news, clickbait, and clickbait. Only one of these sites even bothered to select a Facebook-specific stock photo. And one of the stories is two weeks old.
It’s hard to apply editorial discretion at scale. But it’s not that hard — you can license news feeds from Bloomberg, Reuters, Seeking Alpha, even Business Insider (they do clickbait, but it’s news-driven clickbait, not mad-libs).
The truly frustrating thing here is that un-curated news feeds from high-traffic sites are what these bottom-feeding clickbait factories live on. Google could eliminate this kind of content if they choked off the supply of clicks.
They Took Our Portfolios!
Portfolio tracking is an important feature for personal finance sites, because it gives users a reason to come back to that exact site and aimlessly click around. Like an inbox or a newsfeed, the portfolio page always gives you a dopamine hit, even though it almost never contains new information.
Google Finance’s portfolio tool was pretty clean and straightforward, and offered both a way to track a list of tickers and a way to track the weighted performance of a list of stock positions. Google Finance, in other words, could replace visits to a broker’s site.
In 2017, Google eliminated portfolios. They haven’t come back.
And that’s when their competitive advantage flipped — from minimal interface and a responsive site to muscle memory and distribution.
What’s Wrong With Google
This is almost the wrong question to ask. As a business, Google’s doing great. We should be so lucky as to be doing things wrong the way they are.
And yet, the sorry state of Google Finance says something about the company.
In the early days, one of Google’s competitive advantages was that it didn’t need users to stick around. Yahoo and Altavista thought that search was a tool to get people to hang out on a homepage, where the real money from banner ads would pay for the cost of hosting search. Once Google decided to prioritize getting people off Google.com and to their destination as quickly as possible, they scooped up market share, and then when they figured out how to make money in search, they made a lot.
But the economic logic of the portals wasn’t wrong, just early. If you’re already getting clicks on one web property, it makes sense to expand to another. Most people use few sites and apps, which means that acquiring users is expensive but maintaining them is cheap. This tells you that once you have a user, you should find more things for them to do, so other companies that offer those services never even make it onto the user’s shortlist.
One company that’s done this to amazing effect in a different business entirely is Disney. Their parks and resorts are a more literal sort of destination, but, like a portal company, they realized that it’s much easier to keep someone there a little longer than to get someone new to show up. What Disney pursues is a strategy the parks business calls sequestration: ensuring that if someone is in Orlando for five days, there are five days’ worth of Disney things to do, and ideally enough stuff left over on the to-do list that next year’s trip is six.
They look at demographic holes in their offering, and find ways to fill them. Galaxy’s Edge wouldn’t work as a standalone experience, but as a way to make sure seven-to-twelve-year-old boys aren’t whining by the end of the trip, it’s a winner. (Or, more realistically, to make sure parents who were seven-to-twelve years old during the release or re-release of key Star Wars IP — young enough to form tastes, but not old enough to have taste — can look forward to something that will make them feel ten again. Search your feelings. You know it to be true.)
Google is an advertising company, which means the two raw commodities it works with are attention and information. To put it in extremely shorthand terms, attention represents how many ad units Google can sell, and information determines how much they can charge. Google is a phenomenal success on the information front, which you can see by reading the complaints from companies who used to benefit from underpriced ads (see recent earnings calls from Expedia and Booking for good examples of this).
On attention, they’re also doing quite well. If you look at their financials, you’ll see ad clicks growing 28% Y/Y in the last quarter, with costs per click down 11%. That’s highly distorted by faster growth in mobile, international, and YouTube, all of which monetize a bit worse than a US desktop search click.
Under the surface, there’s a question: are they doing as much as they could to monopolize people’s time?
Sequestration works well for any product with high switching costs, and muscle memory imposes high switching costs. I could have stopped using Google Finance a long time ago, but it used to be better than other options and when I type “f” into Chrome it’s the first thing that shows up.
I can think of two explanations for why Google Finance as a product is not so hot:
- Google is afraid to dominate new categories. Tech companies are stuck in a politically difficult position. Republicans think they tried to swing 2016 to Hillary, and Democrats largely think they should have. This makes them highly sensitive to antitrust, or even the perception of competitive aggression. Ironically, this encourages them to extend their existing monopolies; if YouTube or Google Search is #1 by a bigger margin than last year, it’s hard to tell. But it means that if they go from #5 in a field to #2, they’ve stomped on a lot of competitors who, if they don’t go straight to D.C., will have a quick layover at the NYT newsroom first.
- Google promotes people for launching products but doesn’t fire people for shutting them down. This has been a longstanding complaint about Google, and it has some explanatory value. Reader existed because someone at Google built it, but once they built it they got promoted and didn’t have the time or political capital to keep it alive. I’m told this is no longer such a big deal, but Finance does have the feel of something built in 2006 and given only desultory updates thereafter.
Why Beat Up On Them?
I’ve complained in the past about tech journalists being pointlessly negative. They hear about something new and cool, and stretch to find an argument that it’s old and lame.
I don’t want to just hassle Google for producing a half-baked product. I do want to hassle them for producing a half-baked product, integrating it into Search, which is a billion-plus user product, and then making it worse over time. Distribution gives you some modest moral obligations; if you can require users to use your version of something, you’d better make it good. Microsoft was rightly shamed for making Internet Explorer both crappy and ubiquitous. Today, there’s a new halfheartedly monopolistic bundler in town.
Shaming works for moral failures, not for failures of ability. If Google merely hired lots of incompetent programmers, and Google Finance was their best effort, we could leave it alone. Clearly not the case here; Google still has a high bar for hiring, and Finance is quite obviously the result of smart people not giving the product much attention.
In my personal experience, shaming doesn’t change people’s underlying beliefs, but it’s an extraordinary tool for getting people to behave in line with their views.
That’s my hope today.
If just one person working on Google Finance gets sick of people forwarding them a link to this post, it will make the world better place.