There are three things that really irritate me about A/B testing. The first is where people fool themselves by drawing conclusions from too little data. The second is the myth that small changes frequently result in large improvements and the final one is when A/B tests are used to predict an actual percentage improvement when the data just isn’t there.

You Need a Lot of Data

The proposed improvement

We do a lot of A/B testing at WhatClinic.com and we like to think we know a little bit about the topic. We recently ran A/B test where we put a section of instructional text at the top right hand side of the page. After 11,000 tests and 400 conversions it clearly showed that the instructions made a 30% difference. It would have been so easy for us to stop there and pop open the champagne and boast about how changing one little thing improved our bottom line by 30%.

Things look Great - 30% Improvement

But we didn’t, we kept the test running, because experience has told us not to draw conclusions too quickly.  We let the test run on for another 90,000 people and 3,000 conversion and you know what. In the end it turns out that there was no substantial difference between the two. That’s right no difference.

The whole point of A/B testing is to learn. Learn what works and what doesn’t work. If you don’t run your tests over a large enough sample size then there is a good chance you are going to learn a fallacy. Not only won’t you be moving forward but you will actually be moving backwards and decreasing the value of your company.

Where did my 30% improvement go?

So what if you don’t have the traffic to do A/B tests? Well don’t do them. Do user testing. Get people in and ask them to use your product. You’re going to get a lot more information a lot faster and have a higher degree of confidence in the results.

Small Tweaks rarely makes Substantial Differences

I read about these all the time. You know the type of story – “I changed the colour of a button and increased conversion by 25%”. They read great and play into a pleasant dream that riches and fortunes are just a colour change away. However, in my experience small tweaks have never made a substantial difference to conversion.

It should come as no surprise to you that in order to substantially change user behaviour you need a substantial change to the site. This doesn’t mean that it never happens. However, I suspect that it happens rarely and the bulk of the time it is reported on blog and forums that it is the result of drawing conclusions from too little data or just plain old link baiting. Unfortunately the truth is normally all too boring.

A/B test don’t tell you how much better one page will be over another page

A really common misconception is to think that A/B testing can show you how much better one version of a page will perform over a different version of the page. IT CANNOT. A/B testing can only give you a confidence rate of whether one page is better than another and the observed historic improvement.

Highly advanced A/B testing can tell you a confidence rating of whether there will be a 5% improvement or a 10% improvement, etc, but it cannot tell you what the actual improvement will be. Too often people are fooled into thinking that just because they have observed a 30% improvement during the test that there will be a 30% improvement in the future. Whereas the actual results of the test is that version A has a 93% chance of being better than version B – note no prediction of how much better

Let me know of any examples you have where A/B test have first shown one thing then the other.  I know James Kennedy from voiceover Ireland has one on his blog here

Correction

It has been pointed out to me that the above example only shows a 20% improvement, not a 30% improvement.  Sorry for the mistake

Sometimes you just get it wrong, and in this instance by God were we wrong.

Letting visitors create a shortlist of clinics they were interested in comparing and contacting was a feature we’d been discussing almost since the beginning of the company. It would come up for discussion every couple of months and eventually we got so fed up talking about it we decided to just try it out.

As features go it was pretty simple. All we needed was a link added to each clinic’s search result, one in their profile, and a page to display the visitor’s chosen shortlist. We thought it was a good idea that would be useful to our visitors.

This obviously wasn’t an original idea, and there were plenty of sites that had implemented similar features we could look to for some inspiration. Kayak.com and HostelWorld.com both had directly comparable features, and we also looked at shopping carts on ecommerce sites as well as light boxes on photograph websites.

We planned and rolled out the feature within a week and sat back waiting to see a truckload of visitors happily creating shortlists and coming back to them to compare their options.

That wasn’t exactly what happened.

The Result

We were able to get a pretty good picture in a very short space of time as we have a lot of traffic coming to site. Things weren’t looking good after an hour, and after a full day it became clear that close to no one was using the feature.

We’d had 22,000 visitors over the course of that 24 hour period and only 80 people (0.3%) had added any clinics to their short list. To make matters worse only 17 of those people (0.08%) had subsequently gone back and viewed their short list.

We decided to enter EMERGENCY FEATURE RESCUE MODE. Everyone in the office had their own opinion as to why it wasn’t working. These included:

• The call to action – ‘Add to Shortlist’ wasn’t immediately understandable
• People could not see the call to action link
• It was a crap feature

We changed the call to action text to “Save this Clinic”, pushed it out and waited…

101 people added a clinic to their shortlist on day two. Now normally an uplift of 25% in usage is a cause for celebration. However, when only 0.4% of your visitors want to use a feature it can only mean one thing: turn it off.

So that’s what we did. Two days after launching a feature that we had great hopes for it ended up in the trash can.

Sometimes you just have to work harder on a feature, refining it over time to increase its usage, and other times you just have to accept that you were wrong in the first place and bin the idea. In this instance we figured that no end of finessing was going to create a feature that resonated with our visitors.

Not All In Vain

Over the last three years we’ve gotten pretty used to the idea of launching features that don’t get adopted. A lot of people would look at the effort that we put into these feature as a waste of time but that is not the way we think about it.

Every feature that we decide to develop we regard as a learning exercise. The purpose isn’t to create a fantastic feature; the purpose is to learn something new about our visitors. If you build something and it doesn’t add to the overall knowledge of the company then you’ve missed the lion’s share of the value.

To make this possible we have learned code as lightweight as possible and our processes are now fairly efficient. When we roll out a new feature now we try to expend as little effort as possible getting it to the stage that we can test whether the basic premise is viable or not. Once we decide it is viable we go ahead and refine and improve it.

What was the worst feature you ever rolled out, and how long did you leave it run for? Share your experiences in the comments below.

Every day in RevaHealth.com we talk to clinics that seem to be determined to destroy their own business. They do it the same way that thousands of businesses all over the world do it every year: by not selling to new customers.

When these clinics are contacted by a potential new patient they take the easy way out and send an automatic email reply. Some of them even go as far as sending a personalised follow up email, but neither of these actions is sufficient. To sell to these new customers you have to pick up the phone and call them.

Recently we looked at data from literally hundreds of thousands of email enquiries that have come in through RevaHealth.com and we can say definitively that if your clinic relies on email alone to convert your patient enquiries into paying customers you are cutting your new business by 80%.

Take for example one particular clinic that received over 400 enquiries over the course of a year, and never called any of the patients. Not surprisingly they had a very low conversion rate and weren’t particularly happy with us. Unfortunately they didn’t heed our advice to call the patents and relied on email alone, and in the process they left hundreds of thousands of Euro on the table.

Failing to call your new customers will damn your business and wastes the customer’s time.

Now imagine what your business would look like if you could increase the amount of new business that you get by 500%.

It’s easy to understand why some clinics don’t like to call potential patients. It can be difficult to get your staff to do it promptly, and it is time consuming, but when you realise how much value it can unlock, there is no question but that it has to be done.

When you get an online enquiry into your inbox the natural inclination is the reply to it immediately. Don’t. Take your time read the customers requirement and research the information that you need and CALL them. That’s right don’t email them back, call them.

The process should be:

1. Send an automated response acknowledging receipt of the enquiry and letting the patient know when to expect a call from your clinic.
2. Research their enquiry – potential treatment options, prices, length of treatment etc.
3. Call Them. If you don’t get through leave a message and call them back. Never rely on email alone unless the customer has specifically asked not to be called.

Pick Up The Phone!

Now for the even more surprising finding: nothing else matter as much to the patient as calling them. Not your prices, not your location, not your facilities, not even your qualification has as much of an influence on your ability to convert an enquiry into a customer. They all help, but if you don’t do the customer the simple courtesy of calling them they will not consider going to your clinic.

Happy Customers

This is not something that is still up for discussion. IT IS A FACT. If you want to convert enquiry to customers you must call them. Otherwise you are wasting everyone’s time. A patient will never come to your clinic without you calling them. They will go to the clinic that calls them back.

We spend a lot of time optimising our page load times because we have found a direct correlation between how fast a page displays in a browser and its bounce rate. There are still plenty of improvements left to roll out here, but I was interested in seeing how other Irish websites are faring at optimising their page load times.

I took the companies on the ISEQ 20 Index and added the top 5 indigenous Irish websites based on their traffic from Ireland according to Alexa and tested them for speed.  All tests were run using webpagetest.org and were run from Gloucester in the UK 5 times consecutively. The test emulated IE7 running over a 1.5 Mb/sec DSL connection.  For interest’s sake we also ran it on our own site and saw that we had several areas where we could improve.

A quick caveat: these results were run during just one period of one day and it could easily have been during a bad/busy period for one or more of the sites in question. If you’re interested in making a definitive list of the fastest and slowest Irish sites you should run the test again at various times of the day and on different days of the week.

The Fastest
Unsurprisingly the fastest sites are dominated by relatively simple sites that have largely static content. The page size is relatively low in all cases, maxing out at 250 KB with C&C Group.

1. Greencore 2.0s
2. Grafton PLC 2.1s
3. Kerry Group 2.2s
4. BOI 2.6s
5. C&C Group 2.8s

The Slowest

1. The Independent 15.9s, requiring 165 files and 1,500 KB
2. RTE 9.8s, requiring 100 files and 478 KB
3. Irish Times 8.7s, requiring 127 files and 866 KB
4. Kingspan 9.3s, requiring 91 files and 1,316 KB
5. FDB 9.1s, requiring 72 files and 613 KB

Full Results

In general the results were pretty disappointing with an average full page load time of over six seconds. Some very obvious improvements can be made. Text compression was not enabled on 18 of the 25 sites. There really is no excuse for a website not to have compression enabled these days as it is simple, well understood and implemented on every web server. This can reduce  bandwidth requirements by up to 70%.

The second major area for improvement is the sheer number of files that many of these sites require users to download. Each file has it own overhead and browsers can only download a certain number of files in parallel. On average each of the sites required 57 separate files to be downloaded, with The Independent requiring a truly massive 165 different files. This could largely be resolved by using CSS sprites and combining javascript & CSS files.

Start Rendering Sooner

Of course total download time doesn’t tell the whole story. Sometimes a good user experience can be achieved even if your page takes a long time to finish loading. The way to do this is to get the page to start rendering before the final files are down. The average start to render time for our sample list was 2.7 seconds and this is where some of the Irish web companies shine. For example, Paddy Power and boards.ie, despite being complex sites, start to render after just 1.5 seconds each giving their users a very acceptable experience.

So how does RevaHealth.com compare? Pretty well actually, but with  some obvious room for improvement. We had the fastest overall load time at 1.4 seconds, however our start to render time of 1.1 seconds was 300ms off of the mark. This comparison of home pages may be slightly unfair as RevaHealth.com’s home page is relatively simple compared to one of our typical landing pages. I reran the tests on Dentists in Manchester and we had a full loading time of 5 seconds which isn’t great and needs to be improved. Happily our start to render time was 1.5 seconds which I am pretty pleased with.

Let me know if I’ve made any mistakes in my analysis here. How does your site compare? Post your results in the comments

Over the last month we have been providing local phone numbers for some of our overseas clinics. Essentially we provide them with a UK or Irish number (€240 PA) and proxy the phone calls through to their regular clinic number.  In doing this we have discovered some shocking statistics.

If a clinic’s phone number is looked up by 100 unique individuals it results in 25 unique people trying to call the clinic. You would think that this means the clinic gets to talk to 25 people, but we were shocked to find out that only 45% of the phone calls made result in a conversation with the clinic.

Patient To Clinic Phone Call Statistics

• 30% of all calls are made outside of office hours or during lunch hour
• 10% of calls result in a busy signal
• 15% of calls are not answered
• If there is a voice messaging service available only 30% of people will leave a message
• The average length of a call that is answered is 2 minutes and 20 seconds

Typically a caller who gets a busy signal or no answer will try again immediately and if they don’t get through the second time they give up.  All of this suggests that there must be serious frustration amongst consumer trying to call clinics. We plan on monitoring this area closely over the coming months and will keep you updated on our findings.

Have you had problems getting through to companies by phone? Share your experiences of calling businesses in the comments below.

We provide the health clinics who are our customers with an enquiry management system in order to help them get the best from their patient enquiries, i.e. their customers. We are no different in that we also need a system to manage our customers.

Like a lot of new small companies we were familiar with different CRM solutions, but not enough to know which one would suit us best. We tried an expensive one, a cheap one, and in the end we built our own. Here’s our story.

What Did We Need From A CRM System?

The core functionality that we required included:

• Keeping track of tasks
• Keeping track of leads
• Making notes
• Managing a pipeline

Salesforce

We decided to implement Salesforce pretty early in the company’s life. Of all of the mainstream CRM systems out at the time it was the obvious, albeit expensive choice.

One of the common aspects of a traditional CRM system is that it attempts to model a typical sales process, i.e. Contacts, Leads & Opportunities. From a young company’s perspective this can seem very attractive as it gives your sales process some shape where perhaps there was none before, as was the case with our own. This idea of structuring your sales process is seductive and it gives the illusion of control.

The problem is that the structure acts as a natural restraint to change. As you move your company forward, you often need to make rapid fire decisions and changes to your initial assumptions. However the structured sales process resists these changes. This is not a good thing. In my opinion traditional CRM systems are best suited to mature companies that have very specific requirements which are very unlikely to change.

The downsides of Salesforce in our experience were:

• It is very expensive
• Its feature set is strongly geared towards licensing
• It can be very slow
• It is extremely complex. It can take days to become competent with certain aspects of the software.

Integration Issues

For us one of the major problems we had with Salesforce was the nature of our own business. RevaHealth.com is a directory of health clinics and our live database contains all of the relevant information about each of our customers. Salesforce became a new database that had to replicate a lot of the same information.

We worked very hard to push all the changes made in the Revahealth.com database out to Salesforce as they happened in an attempt to keep the two synchronised. Needless to say this resulted in numerous problems including:

• Changes made to Salesforce were not pushed to Revahealth.com. This was not an immediate problem but over time the problem grew.
• Salesforce couldn’t hope to deal with the scope and flux of the specific information that Revahealth.com handled and this lead to account managers being forced to work with two systems – Salesforce & RevaHealth.com.
• As with all interconnected systems, service failures happen and the effort required to monitor and to fix these failures was greater than expected.
• Once every couple of months we would decide to push the entire RevaHealth.com database into Salesforce to refresh it and every time we ran into licensing issues.

Eventually the SalesForce database and the RevaHealth.com database diverged to such an extent that the sales team no longer trusted it. And that was the end of it. After a year of effort and about 6 man months of integration, maintenance and support we reverted to pens, notepads and Excel.

This wasn’t as bad as it sounds. Even though we had tens of thousands of free accounts to work with we only had around 200 paying customers. In retrospect it is easy for us to say that there is no way we should have even been looking at a full scale CRM system with that number of paying customers.

Highrise

Towards the middle of 2009 one of our sales staff, Owen Cooney, started looking at other systems and experimented with Highrise from 37 Signals. This made a lot of sense to us as we were already using their Basecamp product for our project management. After a month of evaluation we officially adopted it and were very happy that it answered a lot of the shortcomings of SalesForce:

• It was quick.
• It was simple, although excessively so.
• It did not have the licensing issues of Salesforce.

Having learnt from our previous mistakes about database synchronisation we decided that the Highrise system would only be used for new customers and not for account management. This meant that we avoided any integration issues. It worked for a while and was especially effective at task management. However its inability to deal with existing customers and any form of pipeline management meant that our sales team consistently had to use three different web applications:

• Highrise for new customers
• The RevaHealth.com account administration tool for existing customers
• A home-grown pipeline management tool

Of these the single biggest issue for us was pipeline management. As the number of deals that we were creating increased we found that an inordinate amount of management time had to be spent at the end of every month reconciling the sales recorded in Highrise, those in the pipeline tool and the actual invoices generated by the RevaHealth.com administration tool. Invariably there would be substantial discrepancies. This meant that we couldn’t trust either Highrise or the pipeline tool for management purposes.

So What Did We Do?

In the end there was only one answer that made any sense. We built our own CRM system. From the mistakes of the past we now had a very clear set of requirements, and we avoided any sort of replication, duplication, integration and reconciliation issues by building it on top of our existing database.

It took us about 3 man weeks using one of our developers to get it up and running. We’re not the best designers (graphically speaking) in the world but it didn’t need to be pretty as it was only going to be used by our own sales staff. Much more important was that it had to be fast, which thankfully it is.

Seeing as it looks directly at the source data the pipeline report looks at actual invoices not at our sales team’s interpretation of the invoices. It also has a nice and simple task management tool built in to it, and there is no need for two systems for new customers and existing customers.

Building Your Own CRM System Might Not Be Right For You

It’s now more than two years since we first started down the road of using CRM systems. If we had tried to build one ourselves at the very beginning we’d probably just have another horror story to add to our experiences with Salesforce and Highrise, but now was the right time for us to build our own. We had a clear set of requirements and had all the information we needed in one source, our own database.

For you things might be different. Hopefully by reading about some of the mistakes we made you might identify some issues you could run into. Feel free to ask us about our experiences in the comments below, and share your own CRM success or horror stories too.

Leads are the lifeblood of sales. Nothing arouses quite as much passion amongst sales staff as the quality of their leads. The following Al Pacino clip gives a good, if slightly dramatized version of this passion.

Warning: Lots of Swearing!

How Good Are Your Sales Leads?

Sales leads are a strange beast, and the value that is placed on them depends  both on the company and the sales person in question. Give a sales person that normally generates their own leads a list of purchased leads and somehow they’ll manage to call the guy on the list that died last week. All of a sudden the list of leads is worthless in the sales person’s head.

What Determines The Value Of A Sales Lead?

1. Age. In general the younger the lead the better. Older leads might have already decided against the purchase, or even purchased from your competitor.
2. Rarity. The more people who have the lead the less chance that you are going to sell to them. After all, the customer is probably only going to buy from one source.
3. Relevance. How relevant is the lead to the product or service that you sell? The more relevant the better! In other words if you are selling compact cars you are better off with a lead for small cars than with a general enquiry for cars.
4. Stage in the Buying Cycle. Depending on what you are selling it can be better to be at the beginning or the end of the buying cycle. For example, if you are selling TVs then its probably best to be at the end of the buying cycle. The customer has probably made up their mind to purchase and is shopping around for the best price. In the case of an enterprise class software solution though it is best to be around at the beginning of the cycle when you can still shape the requirements.
5. Effort. The sales person’s effort is by far and away the most important factor when determining the value of a lead. If they don’t value a lead (and therefore don’t put the required effort into closing it) then it becomes a self-fulling prophecy.

Obviously the quality of sales leads vary and there are good sales leads and bad sales lead but the quality of the lead is not the only factor in driving sales. You can give top quality leads to an organization and if they don’t value them then they are going to be worthless. Give the same leads to a company that works them tirelessly and they are gold dust.

The effective value of a sales lead is a combination of the original quality of the lead and the value that is placed on it.

Your sales staff will always value the leads they generate themselves the most, because they cost them they most, i.e. their own time. Next will be the leads that cost the organization the most. Right at the bottom of the pile will be the so called ‘free’ leads, i.e. the leads they never asked for but landed on  their desk from some over eager youngster in marketing.

How Do We Know This?

Not only do we depend on sales leads to drive our own business but we also generate large number of leads for health clinics around the world. What we have found in RevaHealth.com is that the clinics that pay for their online enquiries tend to value them the most too. When we survey potential patients two days after they contact a clinic the higher paying clinics routinely get higher levels of consumer satisfaction.  Unsurprisingly they also get a higher conversion rate.

Satisfaction With Initial Communication & Resulting Conversion Rate

* Note conversion rates are extrapolated from the number of reviews generated as result of patient enquiries through RevaHealth.com

It is interesting to note that there is a 50% increase in the number of potential patients who are happy with the customer service they receive from our free clinics compared to our paying clinics. However, there is a 400% difference in their respective conversion rates.

The implication here is that even though 50% of people who contact a free clinic say they are happy with the communications they receive, ultimately only a small proportion go on to convert because their expectations are just about being met. On the other hand, the paying clinics value their leads far more and put in more effort meeting or exceeding the patients’ expectations leading to a far higher conversion rate.

How Do The Top Clinics Convert More Patients?

The top clinics on RevaHealth.com pay the most per patient enquiry and therefore they value them the most.

• They contact every patient  regardless of the perceived quality of the original enquiry.
• They make a phone call as well as emailing the customer.
• They are prompt.
• They are well prepared for each communication.
• They are consistently ranked by the consumer as having above average customer service.

Conclusion
Unsurprisingly spending time taking care of your potential customers through professional and personal communications results in higher sales. Our statistics shows that there is a potential 400% improvement in conversion rate that can be achieved by improving the level of attention that a company pays to its potential customers. There is also a clear link between how much effort a company puts into its communications and how much they spend to create the sales opportunities in the first place.

Do you prioritise leads from different sources? How do you make sure your sales team put enough effort into all the opportunities they receive? We’d love to hear your thoughts in the comments below.

Trials are a very useful way of demonstrating to reluctant customers that you have something that they will value. A trial allows them to see for themselves what your product or service can deliver without committing to the full purchase price. Used correctly trials can be a tremendously effective way of boosting sales, particularly of products or services that do not yet have a reputation.

For a lot of companies, consumer software companies for example, running a trial is effectively without cost. You simply put your trial software up on a website for download and let whoever wants to trial it do so. However, for other companies, including us in RevaHealth.com, trials have a tremendous cost.

Running a trial here with a potential customer means we have to spend time setting them up and teaching them how to use the system. It also means that the trialling company may receive value that would otherwise flow to one of our existing paying customers. The trialling company is also monitored by an account executive on an ongoing basis. All of this means that running a trial from our perspective is an expensive proposition.

The whole purpose of a trial is for the potential customer to determine if the product or service can deliver value. This means the trial isn’t without cost to them either. They have to devote time and effort to evaluating the offering. If they take the trial and don’t spend any time working with it then the trial was obviously pointless in the first place.

So how can you ensure that the customer is going to give the trial the attention it deserves? You can’t really, but a good starting point is to charge for the trial, and here’s why.

What’s The Worst That Could Happen?

A decade ago I was working for Baltimore Technologies selling Public Key Infrastructure into tier-one banks all over the world. Our software was expensive and complex, costing millions of dollars. In the beginning we thought that simply trialling our software for free with one of the world’s leading financial institutions represented a great success. Later on we found that all our product was doing was sitting gathering dust without anyone actively evaluating it. Obviously the entire trial was pointless.

So we started to charge for trials, and as a result we did much fewer, but the trials that we did were always a success. (A successful trial doesn’t necessarily mean that you make the sale – a successful trial simply uncovers and clearly demonstrates the value your product offers to the customer.) Not only did this mean more sales in the long run but it also resulted in significantly reduced costs.

How Did Paid Trials Reduce Our Costs?

Charging for a trial is a nearly foolproof method of evaluating a customer. It tests their interest in the product, their ability to purchase at all, and demonstrates a commitment in a way that is close to impossible through any other means.

1. It forces your own sales staff to correctly evaluate the customer.
2. Even if a customer isn’t interested, sometimes the path of least resistance is to agree to a free trial. After all, it doesn’t cost them anything. Charging for the trial uncovers this.
3. Charging for the trial uncovers if the customer has any money to spend.
4. It also reveals whether or not the person you are talking to has the authority to sign off on a deal, and so helps to identify the key decision maker.
5. Finally, it helps to ensure that the customer will spend some time and effort making sure that the trial gets the attention it needs. After all, they probably have to justify paying for it in the first place.

At RevaHealth.com we never give free trials of our paid accounts. If someone wants to trial a particular account we will normally accommodate them, but they will always pay, normally for a quarter of a year at a pro rata cost. Ultimately, charging for trials reduces the amount of time we need to spend on customers who would never buy in the end.

Charging for trials isn’t going to work for everyone in every sector, but it does work for us. What are your experiences of offering your products or services on a trial basis?

I was reading this article by Barney Austen on Bloggertone and got to thinking about all the things people waste money on when starting their own businesses. There are tons of things that you think are really important to have or do. Many of them really are not.

Here at RevaHealth.com we were no different to any other start up. We invested time and money into areas that should have been left to much later in our development cycle. Here is a list of areas that on reflection we either should have ignored or that we managed to get away without.

A Logo
We still don’t have a logo and don’t have any plans to get a logo. We just typed out RevaHealth.com in a font that we liked and left it at that.

Stationary
We still don’t have any. It was a bit of a pain getting some things set up without any, like a business account with Vodafone, but we managed without and we are still managing without.

Business Cards
We did without business cards for a year and even after getting them printed I keep forgetting them. If your business isn’t based around meeting people face to face then business cards aren’t one of the first things that you need.

A CRM System
We spent a huge amount of effort in the early days evaluating CRM systems. This was at a time when we could list all our customers on the back of a napkin. Eventually we deployed Salesforce at huge cost. After a year we abandoned it and went to Highrise. This still didn’t do the job (more about this in a later post) and most of the company reverted to spreadsheets and notebooks.

We now have our own home-grown CRM system which works like a dream, but we couldn’t have created it in the beginning as we didn’t know the way our business would be shaped in the future. Fundamentally we would have been significantly better off if we hadn’t looked at any form of CRM until we had hundreds of customers and a well defined business. I think many start-ups would be in the same boat and that CRM systems are not something that you should be looking at during the early stages of a business.

A Sales Process
Unless you have already done business in exactly the same business environment and your business model is established and well understood then you can’t have a sales process. Inventing one prior to this is pointless and akin to throwing darts drunk, blindfolded and standing on one leg.

An Intranet
Most of us have worked in companies where a corporate intranet was badly implemented and used.  When I started RevaHealth.com I was determined that ours would be implemented and used correctly. The truth of the matter is that it simply wasn’t required. For small companies the natural organic methods of sharing information are much more efficient. We do use our intranet effectively but it has a much smaller role than we envisioned.

A Server
I spent a good man week deploying and poorly configuring an internal server. It now sits in the corner gathering dust. We don’t have any active servers in our office anymore – everything is in the cloud: Mail server, development servers, intranet, source control … everything.

A Brand Name
This is probably important, however we gotten away with a fairly bland and meaningless one so far. I’m not entirely convinced that if we changed our one overnight that anyone would notice.

A Finished Product
Luckily this isn’t a mistake that we made. We deployed product as quickly as possible and iterated every week from that point on. However, this is a mistake that I see start-up companies making every day. They think that they have to get the technology correct before they start doing business. Whereas they need to start doing business before they can figure out what the correct technology is.

Legally Binding Terms & Conditions
Legally binding to what? A typically start-up changes its business so often that any agreements are outdated as soon as the paper dries. It’s probably best to start with standardised creative commons terms & conditions until your business firms up.

So, when you started up your own business what did you spend time or money on that you needn’t have?

I was at the Weedle offices last night and spent an enjoyable evening drinking their wine and learning about what their future plans are. For those of you who haven’t yet heard of Weedle, they are an Irish Start up lead by Iain MacDonald who successfully sold his previous company Perlico to Vodafone in 2007.

What Do Weedle Plan On Doing?
Weedle see a fundamental and universal inefficiency in how people find other people with particular skills. Regardless of if you are looking for drummer to join your band or you need a patent attorney the process that people go through today is lengthy, difficult and regularly produces poor results. Weedle is looking to solve this problem by creating a searchable social network focused on the skills that people have.

Weedle Homepage

The social aspect to what they do is critical. They believe that a person’s proximity to your social network is as important as they skills they possess. In fact they will only show reviews of people if the reviewer is someone you already trust. To some degree I buy into this, at least for common skill sets that are likely to be in my extended social circle.

How are they going to do it?

• Users create a profile focusing upon their own skills.
• Users expand their network by inviting people they know. They are incentivised to do so becuase the larger their network on Weedle the more likely that someone who is looking for their skills will find them.
• Weedle create a search engine where anyone can look for a particular skill in a specific location. Although anyone will be able to use the search, engine additional value is provided if you are logged in.

The Business Model
Weedle have grand plans and they are passionate about building a scalable platform that genuinely provides value. Some of the fundamental tenents of what they do is that they will never charge for position and they will never charge for anything that doesn’t add value to the platform as a whole. While I think this is admirable I question whether there is a valid business model that can underpin the platform.

I have previously spoken out about Irish tech startups that don’t have revenue models that hold water. In my view companies in Ireland that think they can grow a user base without thinking seriously about revenue need to check their atlas and see just how far away Ireland is from Silicon Valley. Simply put, Irish companies do not have access to the resources necessary for this strategy and by resources I mean captial, expertise, management team, etc. Even though Weedle has some serious financial firepower behind it, I would still hold the same reservations.

Their only disclosed business model so far is pay per click business advertising and I thought it would be fun to dissect it.

The potential revenue from PPC advertising on a site depends on 3 factors:

1. How much advertisers are willing to pay per click.
   This is a function of the content of the page and how far down the buying cycle the typical visitor is. So an advertiser will pay more for the keyword ‘buy lcd television’ than they will for ‘compare plasma & lcd television’ because the second keyword indicates that the user is still relatively early in the buying cylce.
2. The percentage of visitors that will click on the advert.
   This is largely dependent on how relevant the adverts are to the visitors frame of mind. An example of this is that Google can get pretty good click through rates as their visitors are actively seeking information on the subject matter, whereas Facebook gets poorer click through rates as their visitors are being interrupted from their social networking activities.
3. The number of visitors and the volume of pages that they view.

Price per Click
This is all going to depend on how users use the service. If the platform ends up being used primarily for high end skills (legal, marketing, etc.) then Weedle should be able to get a lot per click; a ballpark figure of €1 would certainly be possible. However, if its primary use turns out to be for low skills or worse yet non-economically centered skills then their potential price per click could be as low as 10 cent.

Click Through Rates & Relevance
It is pretty easy to see how you can get the adverts to be relevant on a subject matter basis. For example, legal firms would advertise against legal skills and landscaping companies against landscaping skills. However, it would seem to be difficult to get them to match on social relevance. What I mean here is that visitors have come to Weedle specifically because they are looking to find people with the skills that they have that they are connected to. This makes advertising companies who are not part of the social network irrelevant and as a consequence click through rates will decrease.

Another concern would be that the better Weedle becomes at providing useful socially connected search results then the worse the non-socially connected advertising model will perform. This creates negative incentive for the company which is not a good thing.

So what sort of click through can Weedle expect on their search results? Very optimistically 5% and pessimistically 0.5%.

Volume of Visitors
How often do you need to find someone with a new skill? Pretty frequently for sure, but not on a daily basis. I would think on average about 6 times a year for the typical person. So if Weedle get a user base of a million people who use it 50% of the time, with average page views of 4 it would give them one million page views per month.

Potential Revenue
Based on a million page views a month with click through rates between 0.5% and 5%, and the cost per click being between €0.1 and €1, the potential advertising revenue is between €50,000 per month at the top end and as little as €500 at the bottom. Clearly this is a model that needs massive scale in order to be successful.

To put this into perspective, if Weedle is as successful as Facebook and expand to 400 million users then I would project the maximum PPC revenue at €20 million per month, which compares to Facebook’s €50 million. So why would Weedle earn less than FaceBook with the same number of users? After all, FaceBook adverts can’t be more relevant? The reason is that FaceBook users view a huge number of pages each and log on nearly everyday. Their average user spends 55 minutes a day on their site.

Conclusions
The guys at Weedle are a smart group that have built a tremendous amount of technology prior to launch and they clearly have to be looking at other revenue opportunities. Areas that I think would be worth exploring are:

1. Providing company search results side by side with near equal emphasis to the people search function. Use a pay per lead or pay per position model.
2. Close the loop by processing payment and providing escrow services for certain appropriate skill sets. (The Elance model.)
3. License their search results to the major search engine. (The Twitter model.)

It’s great to see another Irish company setting out with properly grand ambitions, and we can’t wait to see how things turn out for them.

Update -15th of March 2010?

Weedle have just announce $4million in investment for US expansion. This is exactly the level that they need to be playing at in order to succeed. This is a huge milestone and massive success for a pre-launch company.