# Time To First Reponse - Bug!?

**URL:** https://meta.discourse.org/t/time-to-first-reponse-bug/115081
**Category:** Bug
**Created:** [April 15, 2019, 9:28am UTC](https://meta.discourse.org/t/time-to-first-reponse-bug/115081 "2019-04-15T09:28:33Z")
**Posts on this page:** 5
**Page:** 1

<div class="post-metadata">

### Author: ![jousefm](https://sea3.discourse-cdn.com/meta/user_avatar/meta.discourse.org/jousefm/32/81290_2.png) [@jousefm](https://meta.discourse.org/u/jousefm)
#### Post date: [April 15, 2019, 9:28am UTC](https://meta.discourse.org/t/time-to-first-reponse-bug/115081/1 "2019-04-15T09:28:33Z")

</div>

Hi guys!

I am very suspicious about the time to first response data gathered from the forum, I always have peaks in there which do not make sense at all. We usually answer posts within a 8 hour window or faster, so why does is there always a peak. Am I missing something? Example:

 ![TimeToFirstReponseMarch](https://global.discourse-cdn.com/meta/original/3X/5/4/54e1caf17319a71d8da2e239b2de1ddd9e7ef9bc.png)

Best,

Jousef

---

<div class="post-metadata">

### Author: ![pfaffman](https://sea3.discourse-cdn.com/meta/user_avatar/meta.discourse.org/pfaffman/32/120154_2.png) [@pfaffman](https://meta.discourse.org/u/pfaffman)
#### Post date: [April 15, 2019, 11:48am UTC](https://meta.discourse.org/t/time-to-first-reponse-bug/115081/2 "2019-04-15T11:48:10Z")

</div>

Are there multiple peaks or just that day? Isn’t the scale minutes? Isn’t it likely that someone asked a question in the middle of the night and it went unanswered until morning?

---

<div class="post-metadata">

### Author: ![jousefm](https://sea3.discourse-cdn.com/meta/user_avatar/meta.discourse.org/jousefm/32/81290_2.png) [@jousefm](https://meta.discourse.org/u/jousefm)
#### Post date: [April 15, 2019, 12:31pm UTC](https://meta.discourse.org/t/time-to-first-reponse-bug/115081/3 "2019-04-15T12:31:17Z")

</div>

No it is only 1-3 days that have a peak and I take the time to first response **per day**!

Could be but how would you get the 200 hour peak then? 🙂

---

<div class="post-metadata">

### Author: ![pfaffman](https://sea3.discourse-cdn.com/meta/user_avatar/meta.discourse.org/pfaffman/32/120154_2.png) [@pfaffman](https://meta.discourse.org/u/pfaffman)
#### Post date: [April 15, 2019, 2:22pm UTC](https://meta.discourse.org/t/time-to-first-reponse-bug/115081/4 "2019-04-15T14:22:05Z")

</div>

OH. The scale is hours, not minutes. I wasn’t sure about that.

My best guess is that there were some very old topics that somehow had not had responses, like some topic in a category that you’d not expect to get responses. It seems improbable that the metrics are computed incorrectly.

---

<div class="post-metadata">

### Author: ![jousefm](https://sea3.discourse-cdn.com/meta/user_avatar/meta.discourse.org/jousefm/32/81290_2.png) [@jousefm](https://meta.discourse.org/u/jousefm)
#### Post date: [April 15, 2019, 3:47pm UTC](https://meta.discourse.org/t/time-to-first-reponse-bug/115081/5 "2019-04-15T15:47:29Z")

</div>

Thanks for your answer Jay! Already investigated this and I assume it takes also posts from the past which is a bit “odd”.

From what I have found the algorithm looks like this:

`builder.where(“p.user_id != t.user_id”)`

→ Also means that the owner of the topic posting the first reply, should not be considered a reply at all

* * *

Maybe someone else has an idea on this 🙂
