Navigation for Wise Scout - A blog about our online adventures while creating our job board

We have now launched a new job site for the motor trade at www.startmonday.net

 

Thank you for those people who emailed about our beta trials, we still have spaces so please do get in touch. We also have a new holding page for the accounts application here http://www.wiseraccounts.com

 

We are soon to launch a new accounts product aimed at the contractor, freelance and temporary staff markets that makes it really simple to securely store all of your accounts info online and can be easily shared with your accountant. It automates lots of the time consuming and repetitive parts of maintaining your records and is designed to be completely different to Sage and Quicken, which in our opinion do just too much for what we need as contractors. We are looking for some beta testers to give the site a spin and in return will offer 12 months free use of the site after we go live, if you are interested in being involved please email: wisescout@googlemail.com

 

It is not a myth, we have stats to prove it. Yesterday we set a new daily record at 828 visitors, showing that finding a new or first contract is a priority after the New Year.

 

Apologies to any site visitors who have been experiencing timeouts over the past week.  It appears that having in excess of 1 million jobs in our database over the last 3 months has confused it slightly, we think the problem is related to the way it indexes our keywords. We think we have fixed it but only time will tell.

 

Many people who use any of the Host Europe services would have experienced problems accessing our site (and others) over the weekend due to network routing problems on their systems. We had a little panic for a moment when the number of users dropped but now we have investigated the problem we are happy it wasn't with our host. The volume of visitors has picked up again so most of the issues at Host Europe must have been fixed.

 

Last week we reached a new day record of 384 unique visitors, 96% from Google. Not bad for a site that is 3 months old with only £30 spent on marketing. In fact if you see our previous post at the beginning we had 299 in 2 weeks when we first started.
 

When checking how many pages you have indexed by Google put this in your Google search "site:www.wisescout.com" but also check without the www eg "site:wisescout.com". I hadn’t been checking this second search and was getting quite concerned (see previous post) why our listings where swinging up and down so much when we have at least 35000 genuinely unique pages. Google is now showing us with about 9500 pages indexed which is a great relief.

 

The wonder and mystery of Google has perplexed me again this week. While working on our sitemap I managed to get some more pages listed in our map, within an hour Google Bot had been by and starting to index pages quickly, very quickly (the server got a little bit upset and threw a few timeouts). The result was a jump of indexed pages, in fact it doubled over night to 10500. If that were the end of the story that would be great. It wasn’t and within a day the indexed pages dropped down to 4500, over a 1000 less than we started with. Two days later and every few hours we go up and down between 4500 and 5500.  I presume it must be something to do with duplicated content, as our pages are search results of jobs there is a fair chance that some pages will contain similar results, so the next job is to work out how to get Google Bot to visit more unique pages.

 

We have PR at last. The gods of Google have looked kindly on us and have given the main site (.com) a PR of 4 and the blog (.net) a PR of 3. We are very pleased. It has been nearly 3 months since we launched and really had no idea of what rank we would receive. We have over 400 in bound links, some from PR5 and PR6 sites, that we have been tirelessly building up. In fact outside of development it is the most time consuming job. We have also seen a big jump of indexed pages which is now up at just over 4500. We actually currently have over 50000 pages on the site so we hope that Google will continue to index more each week.

 

New US site is on it's way, now we will have a clear UK and US offering, the new site is currently under going testing and we hope to have it live by the end of October. We have received a lot of visitors from the US and we will now be able to offer better job coverage and search facilities.

 

In our world a signup is someone who completes our registration form and by doing so create the basics of their Wise Scout profile. Signups are key to the progress of WS as they establish the core users of the site and begin to develop the real networked nature of our offer to contractors, where they can find and connect with other people they know where having really targeted job alerts and a way of advertising themselves all in one place. But visitors are what we need to get signups, fortunately we have really high conversion rate of visitors to signups so getting people to the site is our top priority. That’s why we have spent a lot of time recently to attract more people to the site.

 

You wont have noticed any new features for the last few weeks as we have been spending a lot of time "under the hood" to work on a number of things:

1. creating some database searches that enable us to send emails to people who fit certain criteria (eg, haven’t added any job alerts, haven’t logged in for a while etc)

2. Making our pages more search engine friendly so they will appear higher up and more frequently

3. Working on our advertising system so we can begin to offer customised advertising by keywords. This makes it much easier for advertisers to focus their adverts to their chosen audience.

 

Having a dynamically created site with most pages having the same base URL with search parameters on them it has been quite hard to get Google to find any of our pages, that is until now. We have finally begun to see our search pages now appearing on Google some very high up for some niche skills. This has happened over the last couple of days and we have seen a clear uplift in traffic to the site because of it.

 

Yes, after a few weeks of heavy development we now have our jobs displayed on maps. Simply do a search and the click "View Results on map" to show all matches on the map, it supports zooming and panning as most online maps. It also has drill through capability on each location to either show the map with your selected results or a list of jobs at the location.

We have also spent a lot of time improving the page loading times and you should now see at least a 40% reduction on all page loading times. We have achieved this through reductions in our css and JavaScript files and enabling some additional caching (especially for images).

 

It has been a busy time for the site, still more work to do on speed improvements, through lots of analysis our page delays (at least 2 secs of them) are caused by our location/mapping link so we need to do some work on reducing that. While working on that we are also implementing our map of jobs feature so that all of the user's searches can be displayed on a map with drill through capability.

A few statistics for the last 2 weeks:
- 299 site visitors
- 1030 page views
- 5:58 average time on the site
- 49% of all visits are typing the url in (or from their favourites)
- 14% are coming from Ecademy clicks

We are pretty pleased with that, obviously the more visitors the better and we have spent a grand total of £30 on advertising so far (which will only really help us with search traffic).

 

We are just testing an new little addition to the login so it will remember your username, nothing life changing but it was a request we received.

We now have passed the 35000 job mark and are looking at how we get up to 50000 over the next couple of weeks.

 

A week since launch and we are ready to roll out some changes to the site:

- page loading dramatically reduced with lots of compression wizardry

- improved search facility that is both quicker and more accurate

- email issues resolved so job alerts should find their way through the more spam sensitive sites

 

We are now working on our next set of features and adding more job sources to our job search to work up towards 30000 jobs a day on the site.

 

Tried a few forums to get a few more eyes on the site and hopefully be able to convert a few more people into signups, got a few responses and one even posted an error we had with some of our search results (it was showing a few salaried jobs rather than pure contract jobs) which was great news and that we have hopefully fixed now. Note the word hopefully as we are never 100% sure that we can catch all the permutations of rates but are actively working on making it better. More details on the search progress to come.

 

After a quick email to a few people I know in IT we pushed our little project out into the world. Thanks to those who tried it pretty much straight away and were quick to tell us we had a few problems with our registration process. Firstly we would have intermittent failures with our emailing sending resulting in a nasty yellow screen of death, the second is that certain hosts appear to be treating our mail as SPAM particularly Hotmail. As a short term fix we have modified the registration process to offer a one step register and sign in, hopefully that won't impact us in the long term but does create a bit more risk that people may not have correct email addresses and therefore may not receive our alert emails but it is a risk we have to take.

 

Our first user was not an internet novice by any means but not a regular user of job sites. After a bit of discussion around the user putting themselves in the shoes of a contractor, understanding the general mindset of how we find work and what we are looking for from a job site. I set out a few tasks that I would like them to achieve. I set them up with a laptop, webcam and Camtasia running (with audio to allow them to comment on parts that were confusing). Pressed record and left them to it. After about 20 mins they had tried pretty much all features of the site. Then I watched the video back which proved to be by far the best test yet, lots of really small details that in hindsight were just plain confusing to the user, some pages that desperately needed additional text to explain what was going on and some combinations of button and key presses we had never tried resulting in some lovely yellow screens.

 

Why we built the recruiter job posting features straight away I will never know (actually that was my fault). It just wasn't needed to launch the site. I need lots of contractors to use the site to be able to attract recruiters to spend their time putting jobs on the site. We should have saved this part of the development and pushed the site live sooner. I hope this isnt a decision we live to regret? We are going to finish them off anyway for now but I bet they don't get used for a while.

So a tip for anyone else, concentrate on the absolute minimum you can get away with delivering, get it live and then carry on.

 

Crikey, what a weekend, I must have spent about 20 hrs this weekend testing. Have raised over 200 bugs/changes from full blown YSOD (yellow screen of death) to simple text changes. 

 

What we have used so far to make this little enterprise:

- Gtalk - free
- Gmail - free
- Visual Studio - coding of site (ASP.NET)
- SQL Server 2000 - database server
- Crossloop - free - desktop sharing
- Fring - free - for gtalk on mobile devices
- Dreamweaver - css layout and design
- Fireworks - image editing

- OS - XP, Vista and Server 2003

I have to also mention that for small companies like us the costs of the Microsoft products would be well out of our reach without the great Empower program. Microsoft are really encouraging small companies to use their tools and products for a really low cost. I think it cost us about £250+ VAT for nearly every Microsoft product including Windows and Office products for 10 users.

 

Everyone can estimate to some degree. When planning more agile like development I use a little process that has worked for me in the past, I don’t profess to be an expert on the area but they may be useful to you

 

A. Ask each team member to estimate their next task ahead, just focusing on the functions you need for this deliverable. Note: don't start talking about your possible plans in 12 months, they don't need to know, you have the strategic view of where its going they work on the now, the tactical aspects of the jobs, if you feel the need to perhaps agile development is the wrong way for you, hop onto a more waterfall based approach and get your architecture sorted up front to support all of the possibilities.

 

B. Monitor how long it actually took.

 

C. repeat a and b until you get the personal multiplier, this is a measure of how much they over or under estimate by, this gives you a stake in the ground.

 

D. optionally apply your own difficulty multiplier feeling if you have enough understanding of the problem/solution domain.

 

Be aware over time each person may automatically adjust their estimates to bring them in line with what they can actually achieve, this experience is usually within the project they are working. So keep a record of this.

 

The site will be based around the Freemium model. Offer a useful set of features to all users for free and then additional features that are worth a financial commitment from those who require them. The features suit the regular professional recruiters and larger employers who will be using the site regularly. The free features are designed for the occasional recruiter who needs a simple way of getting jobs seen. I have always had a target figure in my head for numbers of users and expected revenue, but they are all shots in the dark for now. I need to carry out a bit of research into the value of these additional services. In addition to this we will use a few different types of ads, starting with Google Ads.

 

More work on the paper prototypes, took yesterdays ideas and created a tidied up version. Scanned into a PDF and sent on to development. This will aid much clearer conversations than the text spec I put together. We should be able to get cracking on some of the harder location based parts of the application now we know how it should look.

 

Started the first iteration of the profile pages for the contractor users, paper prototyping is the only quick and effective way of getting the design right. Lots of people have gone to great depths to explain the concept so I won’t repeat their work here. One part I never hear people talk about is the ability to capture the drawings directly onto a computer using a desktop tablet or tablet pc. I have used an HP TC1100 (ebay has quite a few under £500) for about two years and used the tablet capability quite a lot, mostly when capturing ideas and drawings of things I know I want to share with other people. One Note  is consistently good for this job and really does a sterling job at converting my messy handwriting. The export to mht option grabs all of the bits together so you can send it on as one file, some email clients don't like mht but you can zip it first. The other option which I used for this project was to draw and scan the interfaces out, drawing the multiple states of the page where necessary.

 

What a pain, firstly thinking of the bloody thing and then checking that no one else has registered it already, a quick digest of the process of elimination

-           quick keyword game (job, contract, search, location, smart, clever etc.)

-           quick look through domain sites for available names that could be more generic (sedo, sitepoint, dnforum), lots about but none quite right

-           joining names together from keyword list plus some others that became more obvious as I went along

-           checking the names in search of available .coms, even though based in the uk, there is little point in just getting the .co.uk as we have big aspirations for global domination :) and people always guess the .com anyway

And the winner was Wise Scout.

 

Ran the idea past a friend who has been in recruitment for the last 8 years or so. He quite rightly pointed out that they had something similar internal to their company but even this didn't match across the parameters we will put in our system. Quite motivating really.

Got a list of the hard parts of the system to get chewed into, in isolation they all appear manageable but the location based matching looks like  it is going to be the most difficult.

Cracked the initial page by page brief of all the functions we need and where they go. It is going to be quite hard to show any visual differentiation from our competitors but I hope the way ours works will give us something different from our competition.

 

To create a recruitment site for contractors (initially UK only) that adds some additional functionality that I haven't been able to find in any site I have visited, this is truly scratching my own itch as I am a contractor myself. The key problems I wanted to solve:

A. to be able to search one site for all contract specific jobs
B. to be able to create location and rate based searches to take into account the fact contractors usually need different minimum day rates depending on where they work
C. to keep track of where the other contractors are working but more importantly when they are looking for their next contract

I know some of these are covered in some forms on other sites but I hope I can make a site that does all of this and hopefully I wont be the only one who wants to use it.