What put you off my tool?

julian · Feb 25, 2014

Good Evening.

I noticed quite a few of you registered on my 'hobby' www.domtastic.co.uk recently but did not run a valuation and I was wondering why..

(I just want to point out it is still in development and also I'm not trying to be an estibot either.)

So, was this because you thought it was not worth doing or because of the t&c's say I can do whatever with data

or something else?

Maybe you lot can offer some advice on the best way forward with domtastic (scrap it I know

) .

Here's my conundrum.. I have X users registering each day so obviously there is a niche market. Believe or not people do pay for appraisal credits (not so many) so does one aim it at newb's and just open in it up as a free tool where the deal is you get free appraisals of any domain but it is public, which drives more traffic, bit of adsense and commison from domain registrar's

Can't see much point in aiming it at domainer's as they already 'know' about a domains worth.

Your comments and suggestions.

greg2013 · Feb 26, 2014

Ok, some honest feedback, no offence meant at all.

Perhaps run Analytics and see which page gets the most bounce. Might point at why people leave.

I just had a go and I found myself not really knowing what to answer in the drop downs.
While I'm not an experienced domain seller at all, I imagine those who would know what to accurately choose wouldn't really need the form anyway. (?)

So is this aimed at pros (who wouldn't need it?) or noobs who cannot know what to choose correctly?

Spacing the drop downs apart a little might help, as their being squashed up together makes it a bit "cluttered".

When I tried, you reported:

RESELLER VALUE £1,100/£1,250+
POTENTIAL END USER VALUE £11,500/£12,500+

So being a noob and not really knowing what to choose in the drop downs, I tried again and changed one drop down only:
"Does this domain contain keywords or phrases that relate exactly to its potential market area"
From "Yes" to "Yes exact keyword/phrase match"

You then returned:

RESELLER VALUE £2,000/£2,500+
POTENTIAL END USER VALUE £22,000+

Twice as much? So with all other criteria considered, my domain being from "matches keywords or phrases" to "exact keyword or phrase match" gives me £22k rather than £11k?

Just doesn't seem like it's considering enough things.

But last of all, and most importantly (potentially why others have not returned):
You emailed me my plain text password. Sigh!

Just don't.

I used a new one that I'll never use on anything else, fortunately, but if I hadn't I'd have been sorely annoyed right now that you stored my favourite password on your site for any database hacker to get, and sent it around the UK network of DNS servers for any chump sniffing networks to discover it.

Even beyond all that. I don't want anyone seeing my password (and this includes you looking through your database)

stellar73 · Feb 26, 2014

You've answered your own question when you said it wouldn't appeal to domainers as they "know" the value of domains.

In other words the tool (as with other automated evaluation tools) attempts to solve a problem that doesn't really exist.

Admittedly the "problem" may exist temporarily for newbies/non-domainers but even then you're probably going to get people trying it once or twice for novelty value and then moving on (especially if and when it dawns on them that the only real value of a domain is what a buyer is prepared to pay for it -and seller is prepared to let it go for).

If you are going to keep going with it then making it free supported with adsense sounds like a better option. I imagine if you're selling credits to people you're opening a potential can of worms for the future, ie the type of people that might spend money on a valuation are probably the same type of people who would direct all their anger at you once they realise they cant sell their domain for £20 never mind £2000.

Edwin · Feb 26, 2014

As far as I can see, the tool doesn't do any sort of analysis on the domain name itself beyond checking whether it has numbers in it or not, i.e. the values returned are dependent on the choices made from the drop downs.

As such, you almost don't need people to type the domain name in at all! (I tried bddritain.co.uk and britain.co.uk with the same drop down choices and got the same value - yes, I know I was fibbing in my drop down choices for the first one, but it showed me there's no word list lookup)

Agree 100% with the comment above about plain text passwords - that's a real no-no.

PoshTiger · Feb 26, 2014

I really liked the description it presented when I asked for the valuation - the reseller price appears to be pretty accurate too... The end user price is a little high though.

I think a 'Rate my domain' type site would be the only accurate way to go, where members put their domains up for valuation and get responses from other members on their value.

I think valuing domains is far too subjective to be automated and needs human input - if nubcakes don't like what they see, they go to other valuation sites until they find a valuation that meet their bonkers expectations.

julian · Feb 26, 2014

Some good points - thank you all for your feedback.

I will change password issue

@poshtiger - Glad you like the valuation and found prices accurate. Rate My domain maybe go.

@Edwin - you cheated and manipulated the system so you got a dodgy valuation

@stellar73 - yes I see this is basically a novelty/newb tool that someone might use a few times and move on. At the very least I hope the form selection will make them think about how they value a domain.

@greg - all good points. The difference between partial match and exact match is significant value indicator hence the increase in valuation. There is nearly no bounce rate - only from AD users

Thanks all..the fact remains people register and use this site every day but looking at the average domains going through are mostly low value. So regardless of whatever 'system' I use or how 'good' it is, it must be fair to surmise my target market is 90% newbs.

bensd · Feb 26, 2014

I'm sure you'll get some feedback about the tool itself however, 99% of domainers aren't going to be interested in an automated valuation IMO

greg2013 · Feb 26, 2014

So you’re happy it's for newbs, then might I suggest focusing on that direction, and adding some details for the drop downs.
Or, maybe make them radio buttons so one can hover a question mark next to each option and read info about what each choice is, with examples etc etc, then select the one most suitable.

And for the passwords, I recommend using Bcrypt and Blowfish (so encryption and random salting). It's pretty straightforward and does most of the work for you once you know the commands to call.

But certainly don't give up on it, just make it fit for purpose. A lot of web services have teething issues and need tweaking here and there.

julian · Feb 26, 2014

I took on board a lot of comments cleaned it up a bit, and binned the user registration thing can't see any point at this stage in making people register. Still needs more work, shit I wish I was a decent coder.

Had about 6,000 valuations so far so let's see what happens.

ta

shameless plug time for free domain valuations

greg2013 said:
So you’re happy it's for newbs, then might I suggest focusing on that direction, and adding some details for the drop downs.
Or, maybe make them radio buttons so one can hover a question mark next to each option and read info about what each choice is, with examples etc etc, then select the one most suitable.

And for the passwords, I recommend using Bcrypt and Blowfish (so encryption and random salting). It's pretty straightforward and does most of the work for you once you know the commands to call.

But certainly don't give up on it, just make it fit for purpose. A lot of web services have teething issues and need tweaking here and there.

Retired_Member42 · Feb 28, 2014

I'm not into these tools at all. I tried carinsurance.co.uk with everything answered correctly and got an £8 valuation. As others have said, it's virtually impossible to devise an algorithm to accurately value domains. Sure some will get an accurate valuation but the vast majority won't. We all know how annoying it is when somebody lists a £10 domain for sale here for four-figures because of an Estibot valuation. The last thing we need is more of them.

I don't think even having a pooled domain valuation site taking on board several opinions will work. Firstly, I think very few will spare the time to accurately value a given domain and even if they did, as we've seen here, how one domainer prices a domain can be vastly different to how another will view it - and that's experienced domainers. I just don't see an angle for domain valuations short of someone who is in the know really doing the research and manual human leg work on a case by case basis.

Not trying to piss on your chips but I just don't think it has legs. I think most of the valuations are people like me throwing in decent names and just seeing what your tool thinks.

Murray · Feb 28, 2014

This is just constructive criticism (well, it's meant to be)

It's impossible for a tool to give an accurate valuation, which you know, so why make one? I don't understand the thought behind it.

Also Domtastic sounds like an adult site.

julian · Feb 28, 2014

Thanks, points taken. The reason carinsurance was so low is because of the length.
The algo likes shorter domain and knocks down 12chars in length big time.

I reduced to 4chrs. and one word and got a 40k valuation. System does not like 2 word long domains - I need to tweek

Redemption said:
I'm not into these tools at all. I tried carinsurance.co.uk with everything answered correctly and got an £8 valuation. As others have said, it's virtually impossible to devise an algorithm to accurately value domains. Sure some will get an accurate valuation but the vast majority won't. We all know how annoying it is when somebody lists a £10 domain for sale here for four-figures because of an Estibot valuation. The last thing we need is more of them.

I don't think even having a pooled domain valuation site taking on board several opinions will work. Firstly, I think very few will spare the time to accurately value a given domain and even if they did, as we've seen here, how one domainer prices a domain can be vastly different to how another will view it - and that's experienced domainers. I just don't see an angle for domain valuations short of someone who is in the know really doing the research and manual human leg work on a case by case basis.

Not trying to piss on your chips but I just don't think it has legs. I think most of the valuations are people like me throwing in decent names and just seeing what your tool thinks.

Retired_Member38 · Feb 28, 2014

I'm amazed someone who has been around as long as you is wasting time and money on an absolute turd of a business model like automated valuations.

Julian, you know yourself the valuations it's throwing out are a complete waste of time. You know it, I know it, everyone reading this thread knows it. No amount of tweaking is ever going to change this, surely you accept that also?

So why bother, what's the end game here? As I can't see it.

PanzerWagon · Feb 28, 2014

Monkey said:
So why bother, what's the end game here? As I can't see it.

No idea but I clicked on the adsense by mistake so all is not lost!

julian · Feb 28, 2014

I don't understand what you mean - can you put it more bluntly?

It's just a hobby really - something I code while I lift my feet for the cleaner

. I coded myself so nada cost.

Sometimes it bangs out (imho) a spot on valuation which I find interesting - other valuations are total shite. I'm under no illusion it's not giving consistent, robust valuations.

There must be a market for domain valuation tools as some days I get double figures + going through it.

I don't know what the end game is yet.. but I'm getting closer.

Monkey said:
I'm amazed someone who has been around as long as you is wasting time and money on an absolute turd of a business model like automated valuations.

Julian, you know yourself the valuations it's throwing out are a complete waste of time. You know it, I know it, everyone reading this thread knows it. No amount of tweaking is ever going to change this, surely you accept that also?

So why bother, what's the end game here? As I can't see it.

Retired_Member38 · Feb 28, 2014

julian said:
Sometimes it bangs out (imho) a spot on valuation which I find interesting - other valuations are total shite. I'm under no illusion it's not giving consistent, robust valuations.

You might get the very occasional correct valuation - but I've got a broken watch that is correct twice a day also

julian said:
There must be a market for domain valuation tools as some days I get double figures + going through it.

I don't think there is. It's a market for suckers.... It's only clueless newbies posting worthless domains along with an £xxx estibot valuation. No legitimate buyer or seller is using any automated valuation service as they understand automating the process will simply never give valid results.

If you're just doing it to have a bit of fun/practice coding or whatever then fair enough! but I can't help but think that time could have been spent building a tool that actually helped you do something of value. IMO an automated valuation of domains is just never going to work.

namealot · Mar 5, 2014

Isn’t a failing in all domain valuation tools (apart from inaccuracy) that there unlikely to convince or make end users any happier to part with the dosh?
Its hard to get around it cost you x you want xxxx ???? Thats always going to hinder even the best of them

greg2013 · Mar 5, 2014

Auto evaluation sites are sometimes 90% accurate, sometimes 10%, sometimes somewhere in between, etc.
This fluctuation in inaccuracy makes them 100% unreliable.
They're already pointless for seasoned domain sellers, and misleading for noobs.

For it to be relatively reliable, and thereby useful, it needs to be around 80-90% accurate at least 85-90% of the time.

I suggest, if you don't already, save all the results in the database - everything the user chose, the domain, and the end prices your script provided.
Have a page in your admin (not public) where you grab all that data from the database and display each domain and the relevant data in a single row in a table (easy to see).

Study it.
Manually reviewing should reveal where the scripts are failing. Tweak for that particular issue, re-assess the data, tweak something else, etc.

These tweaks are also probably going to require additional calcs in the code. ie Don't just tweak your current ABCD checks to incorporate a problem, add additional EFGH checks to all the current IF/ELSE or classes (whatever) you likely have currently, so it's more accurate for each particular scenario, not just a blanket best guess for a range.

Retired_Member38 · Mar 5, 2014

Greg but its worse than that - it will be failing (at least in part) due to data inputted by the user, as they're being asked to answer questions on their own domain. So no amount of tweaking is going to fix it.

I didn't think it was possible, but someone has actually managed to out-turd Estibot

In all seriousness.... I don't think people building/promoting/linking to these things is helping you domainers in any way at all. Look at the completely idiotic threads that pop up every week or so with someone asking stupid money for a domain that is truely worthless, and trying to justify it with an Estibot valuation. Why add to that mess?