Teardown

Five attendance apps, 443 reviews, and the registers nobody could get back

We read the complaint record for five attendance apps on Google Play before deciding how our own attendance program keeps its register. The worst thing an attendance app can do turned out to be the commonest complaint, and most of it traces to one update.

What this is based on 443 Google Play reviews across five attendance apps, first read on 19 September 2026 and re-pulled on 16 October so that every count comes from kept files and a saved script, and the backup after every pull that the data-loss complaints decided.
In this piece
  1. What we read
  2. How the counting works
  3. The commonest complaint is the worst one
  4. Updates and logins, which are the same complaint twice
  5. The phone is a weak place to take attendance
  6. Where a terminal does not save you
  7. What this decided, and what we are not claiming

We make an attendance program, Herewise, so this comparison is not neutral and we are saying so before the first number. It is a Windows program that runs offline against a ZKTeco K40 fingerprint terminal, it is in use in one university department, and it is not on Google Play. None of the apps below is quite what it is. They are what a lecturer or a club secretary finds when they search the store for attendance, and their reviews are the clearest record there is of what people punish an attendance app for.

What follows is not an opinion about which app is best. It is the review record, counted.

What we read

We first read these reviews on 19 September 2026, before deciding how our program stores its register. That run kept a summary and the worst reviews of each app, not the full files, so its counts could not be checked by anyone else. On 16 October we pulled the same five apps again, in English, from the US store, sorted newest rather than by Play’s default relevance, and kept every file. Every number below comes from that second pull.

App Installs Store rating Reviews Window Negative, with text
Attendy 100,000+ 3.73 of 994 201 Jul 2021 to Oct 2026 84
ZKBioTime 100,000+ 3.32 of 517 87 Mar 2018 to Jul 2026 40
Attendance Radar 50,000+ 3.88 of 117 53 Nov 2022 to Oct 2026 6
Qwickly Attendance 10,000+ 1.93 of 154 39 Sep 2022 to Oct 2026 37
InClass 10,000+ 3.58 of 300 63 Aug 2024 to Oct 2026 11

443 reviews, of which 178 are one or two stars carrying text. Only those 178 are counted below.

Look at the windows before the numbers. ZKBioTime, the phone app from the company that makes the K40, needed eight years to produce 87 reviews. Every app here is small next to a Pomodoro timer or a music stats app, and a category this quiet will not give you thousands of reviews however long you wait. The percentages describe a small record, and we would rather say that than dress it up.

How the counting works

Each theme is a set of keyword rules run over the text of the 178 reviews. Some need two things in the same review, a subject and a failure: a review counts as “an update broke it” only if it mentions an update and something going wrong. A review can count under more than one theme, so the shares do not add up to 100. A keyword can misfire, so treat each count as close rather than exact. The script prints every review behind any count, which is how three of the themes below were checked by reading each match.

The commonest complaint is the worst one

Theme Reviews Share of 178 Mostly from
Records lost or wiped 38 21.3% Attendy 37
An update broke it 35 19.7% Attendy 20, ZKBioTime 8, Qwickly 5
Paywall or payment 26 14.6% Attendy 23
Cannot log in, or keeps logging out 21 11.8% ZKBioTime 10, Qwickly 9
No reply to an email or a support request 17 9.6% Attendy 13
Hard to set up, no instructions 15 8.4% ZKBioTime 6, Attendy 6
Slow, crashes, will not open 13 7.3% Attendy 6, ZKBioTime 4
Check-in depends on the phone 13 7.3% ZKBioTime 8, Qwickly 4
Language 6 3.4% Attendy 4, ZKBioTime 2

An attendance register is the one kind of data where losing it cannot be waved away. It is evidence: of who was in the room, of who may sit an exam, of who paid for a term at a club. And the largest theme in the record is people losing it.

It is also mostly one incident. 37 of the 38 are Attendy, and 30 of those 37 name version 3.6.2 or 3.6.3. They start on 31 July 2025 and run until December. A French reviewer dates it exactly, to “votre mise à jour du 30 juillet 2025”, your update of 30 July 2025. So this is not a rate at which attendance apps lose data. It is what one update did to one app’s users, and what they said about it afterwards.

What they said is the useful part. On 3.6.2:

I’ve been using this app for quite a while which has been very helpful in keeping tabs on a sports club I run. Unfortunately (without any warning) all data has now been wiped due to a bizarre update which has completely changed the language and format.

And the one that explains why nothing could be done, also on 3.6.2:

Each time I added to my records, the app said that the data was saved locally. But there is no indication of where it was saved locally that I can use it to restore my data.

That second review is the whole lesson in two sentences. “Saved locally” is true and useless if the person cannot find the file, copy it, or put it back. A backup you cannot see is not a backup you have.

Another reviewer, on 3.6.2, had chosen not to pay for the cloud copy:

I have been using for over a year. I did not use the cloud because for me it was not needed. I logged in last week after a 1 month break and now all my data is gone without warning.

That is where the paywall theme comes from too. 23 of its 26 are Attendy, and 13 of those 23 also describe losing the register: they had paid for premium, or had declined the paid cloud backup, and lost it either way.

One thing in Attendy’s favour, because it is in the record and the reviews alone would hide it. The developer has replied in the store to 158 of the 201 reviews, 78.6%, more than any other app here. Thirteen of its negative reviews say an email or a support request went unanswered, and we read all thirteen. Both are true, and they describe different channels.

Updates and logins, which are the same complaint twice

Updates that broke something are the second theme, and unlike the data loss they are spread across apps. On ZKBioTime 2.0.24:

After updating to the latest version (2.0.24) yesterday, I’m not able to check in/out using the app. The app is not able to detect the location even after granting permission to Location services.

Logging in is the other half. Ten of ZKBioTime’s 40 negative reviews and nine of Qwickly’s 37 are about signing in. For an attendance app that is not a minor irritation, because the sign-in sits between the student and the record. On Qwickly 6.3.1:

It’s anything but quick. takes like 3 minutes to sign in just to scan a QR code, and if you witch to a different app in the middle of it and come back, you have to restart

Qwickly is the starkest record here. 37 of its 39 reviews are one or two stars, and the developer has replied to none of the 39.

The phone is a weak place to take attendance

The rule finds thirteen reviews where the check-in depended on the phone and the phone did not cooperate. Reading them, twelve are that: location not found, a camera that would not work, NFC or a QR code that would not scan, a brand the app did not support. The thirteenth is a screen stuck on a picture. On Qwickly, with no version given:

Can’t use it unless i buy a new phone. Thanks for making my life in university a living hell.

Taking attendance on the student’s own phone also makes the student’s phone the weakest part of the system, and the one the institution has least control over.

Where a terminal does not save you

ZKBioTime is the phone app for BioTime, ZKTeco’s own attendance server. The same company’s desktop software, ZKBio Time.Net, is what our program replaces at one university. ZKBioTime’s reviews are worth reading for anyone with a fingerprint terminal, because they show that a terminal on the wall does not remove the setup problem. It moves it.

New user here, what/where do i get my server ip and port?

That is from version 1.0.5, in 2019. And from 2020, with no version recorded, the terminal we work with:

Not working for me on ZKT K40

We cannot tell from one line why it failed. We can say that getting a laptop to talk to a K40 took us long enough that we wrote a five-step check for it.

What this decided, and what we are not claiming

The record decided two things outright. We read it on 19 September and the automatic backup was committed the next day.

  • It backs the register up by itself after every pull that stores scans, and keeps the newest ten on the same computer. Nobody has to remember.
  • A backup can be put back from inside the program. It shows what the file holds beside what the register holds now, changes nothing until the person confirms with their password, and copies the register as it was first, so a restore can itself be undone. It has its own tests.

Two more were already true of the program, and the record is why they stay.

  • Check-in happens at the terminal, not on a phone, which avoids the phone-dependent cluster, and makes it harder for one student to sign in for another.
  • The program reaches the terminal over a cable without the internet, and walks five checks in order when it cannot, because setup is where the ZKBioTime reviewers gave up.

Four honest limits on everything above.

This is one locale, US and English, and the apps do not all serve the same people: Attendy is used by clubs and teachers, Qwickly by universities with an online learning system, ZKBioTime by employers. A Nigerian lecture hall is in none of these records directly.

The windows are not equal, and the whole corpus is small. Eight years of ZKBioTime reviews are 87 reviews.

The largest theme is one incident at one app. It shows what users punish, not how often it happens.

And the counts come from keyword rules, not a careful reading of all 178. The largest theme, the support theme and the phone theme were checked by reading every match, and the rest were not. The files and the script that counted them are kept, so anyone disagreeing with a number can disagree with something specific.

Written by Adeboye Oluwatimileyin and Moses-Azuoru George, the two people who are Emberfig. We make money from AdMob inside the Android apps and from setting Herewise up for departments that ask. This site carries no ads yet. Using an app never costs you anything, and we never sell your data.