from Hacker News

Traffic spikes are bad for your product

by twapi on 1/31/25, 1:47 PM with 75 comments

• by YmiYugy on 1/31/25, 4:17 PM

  I did not find the article very compelling. The biggest problem is the lack of context. What kind of products is the author talking about? I got the impression that this is mostly about SaaS offerings, but who knows? Moreover, the arguments presented are insufficient with regards to it's thesis. The author claims that traffic spikes are actually harmful, but then just argues that they are mostly irrelevant. The only argument towards actual harmfulness presented is that disappointed users could spread negative sentiment. I don't find that very convincing. Something conspicuously absent from the article is a discussion of conversion cost. If you have a freemium offering how much do free users actually cost you? Especially for consumer facing software the cost can be really low. It's not like you have a sales team doing calls with each potential customer, people crowding your store. Might a surge of signups still be profitable, even if the conversion rate is much worse? But even if you accept that unmitigated spikes are harmful, aren't the remedies easy to apply. If you notice an unwanted spike you can easily increase friction by e.g. limiting new signups to referrals until the spike is over, temporarily restricting your free tier, etc. In conclusion, traffic spikes might not be that harmful and are easy enough mitigated as to not warrant taking action to prevent them.

• by analog31 on 1/31/25, 3:38 PM

  The thing that didn't seem obvious from the article is that "looking" means signing up. If this is the case, it's a conscious choice of the vendor.

  On the other hand, I do know that it's a widespread feeling among salespeople, that a person who interacts with them, without buying anything, is stealing from them. And wasting someone's time in order to make them feel like they owe you something is a familiar bargaining tactic.

• by hartator on 1/31/25, 4:07 PM

  > these shitty users

  God. You’ll never be successful if you are that insulting to your users.

• by Oras on 1/31/25, 4:04 PM

  I agree with most points, however going viral has the benefit of strong backlinks for SEO.

  I had a website I shared here on HN, which made it to the front page and resulted in being mentioned in multiple newsletters and websites. The website site was on Google's 1st page when you search for the term, all due to the strong backlinks I gained.

• by philipwhiuk on 1/31/25, 3:09 PM

  This is true in every field - a stratospheric number 1 kills an artist.

  https://www.youtube.com/watch?v=9LZEZ5QuyzM

• by jmuguy on 1/31/25, 3:08 PM

  IMO you see the same issues when you have Google start optimizing ads for user signup. We've had a huge amount of user growth from ad spend. But not really the users we want, for the most part.

• by cess11 on 1/31/25, 5:03 PM

  From https://a16z.com/andrew-chen/ it seems this person failed at running a company that his rich friends had invested in and then they kept him around as some kind of consultant.

  I'm not so sure it's the kind of person I would turn to for advice on anything.

• by AdmiralAsshat on 1/31/25, 2:38 PM

  > (I learned this phrase a long time ago from a friend who used to sell cars!). When your product goes viral, yes, you do a big spike in users. But what you’re really getting is an Invasion of Looky-Loos, low quality users who come in, check things out, but don’t stick.

  I find it strange that we generally frown upon the "Buy something or get out" mentality in brick-and-mortar stores, yet we seem to be cheering on this mentality if it's an online service. It seems like castigating your potential customers because they didn't give you money immediately or give you as much money as you wanted is a shitty business practice.

• by incognito124 on 1/31/25, 2:37 PM

  There's something icky in calling users high-quality and low-quality

• by dylan604 on 1/31/25, 3:54 PM

  Since the beginning of time (of software released on the interweb), people have been burned by the software not doing what it says on the tin. Or that it sounds like it might do what you really need even if that need isn't advertised on the tin.

  The reality is that people have needs and look for solutions. Some needs never have an off the shelf solution and it might take multiple OTS options strung together with some glue code. You don't know what's possible until you get a better look at the software. Why someone would be amazed that their perfect little product is not actually, you know, perfect solution for everyone is the real tell.

• by TZubiri on 1/31/25, 4:39 PM

  Ngmi

• by mtlynch on 1/31/25, 2:30 PM

  Clickbait title.

  This isn't a story about something that personally happened to the author but a general argument against big launches. They don't offer any concrete examples like the title would imply.

  The article itself is pretty thin and quickly devolves into a bunch of analytics jargon word soup like "Metrics: D1/7/30, consistent weekly growth (5-10% WoW), NPS (and qualitative feedback), DAU/MAU."

  Edit: The original title as submitted to HN was "My product went viral on social media but all I got were these shitty users." It looks like the mods have since changed it to a less misleading title.

• by asdev on 1/31/25, 4:33 PM

  TLDR retention is more important than top of funnel

• by dogma1138 on 1/31/25, 2:25 PM

  “My product went viral but it may not be offering the value I think it does so I blame the users for it”

• by davidee on 1/31/25, 4:41 PM

  "low quality users"? I think you mean folks trying to assess if it's a low quality product.

  I just couldn't finish this thing. Disdain mixed in with some wild understanding of the world.

  The fact that we so often care about DAU/MAU as a key metric for running a business continues to be one of the many issues some of us face when trying to build "quality" businesses of any size/scale.

  "here at wannabeMegaCorp our mission is numberGoUp"

  cool cool I'm out (even if I missed a later point). I'm all for reading a take I don't agree with, but this was something else entirely.

• by jaimsam on 1/31/25, 3:10 PM

  That most "people" on social media are flakes, or looky-loos, NPCs, or just bots shouldn't be news to anyone.

• by sksrbWgbfK on 1/31/25, 2:31 PM

  > you launch an app

  "Free ReactJS web site filled with A/B testing and analytics, and a billion NPM modules" I guess.

• by ChrisMarshallNY on 1/31/25, 2:38 PM

  That's one reason that we personally vet every signup request for the app I run.

  One of the services that we offer, is users being able to (anonymously) reach out to each other for help. If we have "looky-loos," or even scammers, it's not particularly good.

  Scammers usually have no use for our app. It's really too anonymous and locked-down, and our users tend to be fairly poor (and scammer-wise -some are even reformed scammers). Faecesbook is a much richer hunting ground. They may sign in once (if they were able to get us to admit them, which takes a bit of work), and then never come back. Some of them never log in, because they don't actually look at their emails. If they never log in, then we delete the account after 30 days.

  Looky-loos, on the other hand, are a pain in the ass, because they will log in once, peer around, realize the app has nothing for them, and never come back. We need to wait a year to delete their account.

  We absolutely don't care about scale. We may never have more than a few thousand users. We haven't even reached 1,000, and it's been a year. I tested the app with 12,000 fake users, so it can handle that.