Own web analytics for startups – Part III

Part III – thinking about surrounding our product systems

Today we will fill column A (web applications) in our table and make some conclusions about systems needed around a web application product to make it successful.

A1 – Marketing site. Is it needed? Absolutely, how else will potential customers find out what your application does, fall in love with it, and become customers for life? But as we said before, the functions of a website do not end there. It serves as the core, around which all the other systems build up into one whole. I’m not sure if examples are needed, but if you want to see some samples, feel free – JanRain, ChartBeat, Hipmunk. We put a + here.

A2 – Blog. Perhaps you will disagree, but in our opinion blogs are quite useful for developers of web applications. You can use your blog to announce upcoming changes to the product, which will show what you are working on, what direction you are heading in with the product development. On the other hand, it’s a convenient means of sharing some of your thoughts with your current and future customers, dispelling the fog of mystery surrounding product development.

A3 – Forum. The main purpose of a forum is the interaction between people, let’s narrow that down to web applications business; customers between themselves, customers with developers and possibly, potential customers with all of the above. Can’t you see the usefulness of this kind of interaction? We can, and also, a forum can be used as underlying platform for Answers and Ideas systems. Put a +.

A4 – Customer Support. While your product is free it will be enough to just have a Answers portal or a part of your forum dedicated to it. But if your application has a price tag attached, then you have certain responsibilities to your customers, and therefore you need a way for your customers to contact you directly. Some people use a special mailbox, which may work with a small user base. But if you plan to have a large number of paying customers, then it’s better to use a customer support system straight away. It’s a poor soldier that never wants to become a general. Put a + here.

A5 – voting for Ideas. In this series of posts we are talking about start-ups or; the people who have just started, they don’t have thousands of addicted customers who love the product as it is. So it is very important to have a feedback channel established with those who weren’t afraid to become early adopters. Are they happy with everything? What do they need to be totally happy? These questions can only be answered by the customers themselves, so give them that opportunity. On the other hand, you may want to hear what the customers have to say about your planned alterations before you start coding, or do you prefer to find out after a month of hard and maybe pointless work? +

A6 – Answers portal. Let’s put ourselves in the shoes of a web apps developer (we have fit in those shoes perfectly), now tell us honestly, have you ever dreamed of answering the same questions for days on end, asked not even by customers but also by visitors? The Answers system lessens the burden of this task. Can you feel how the fear slowly drains from your body, the fear that was generated by the vision of answering all those questions? We put a +.

Our table is finally filled. And it’s clear, that for web apps to work effectively with its’ customers, practically all of the systems mentioned in our table are needed. Of course, nobody is forcing you to have all these systems from get-go, and you may not need all of them.

So everyone defines a set of systems depending on necessity. But where do you get them? Which systems to use?
It would be awesome to get all of these systems in a neat little package, in one platform, but as it appears, there is no such product, payed for or free, on the market to date. Or at least we were not able to find it. If you know of a product like that, please leave a comment. How great would it be to just select check boxes next to components you are going to use, order only one(!!!) theme fro a designer which will cover all components (like WP) and you’re done. But no, you’re forced to integrate your product with each system individually. It’s a nightmare, don’t you agree? I would gladly pay for a product like that. What about you?

Of course, if someone somewhere makes a product like that, it would be logical to add web analytics to it, for now we will be building something like that ourselves. So there you are working on your product and you end up finding a few ‘fifth elements’, which are not on the market, and for which you would pay for gladly. Each of those ‘elements’ try to switch you over to making it, if you don’t find anything, which would completely suit your need. But we ‘tie ourselves to the mast and sail through these stormy seas’ with thoughts of revisiting on the way back.

Because we are going to have to do everything ourselves, let’s try to decide which combinations of existing systems we should use, and discuss it’s pros and cons.

Marketing site – if you are proficient with any kind of CMS, there are no questions asked. Otherwise, if you take into account the amount of pages and the frequency of those pages being altered, the installation of CMS, its learning, and designing pages for it will take more time than using static HTML or PHP with a templates (Smarty).

Blog – the main candidate here is the omnivorous and free WP. Another blog platform can be considered due to personal preferences, or if it integrates easier with other components.

Forum – there are many of these systems on the web, free (phpBB) and payed for (vBulletin), which can be installed on your server.

Customer support – to stat we would ask our clients to contact us via email. Internally we will be using OTRS or Eventum, issuing a ticket number and rolling forward.

The only systems left are Answers and Ideas.

Answers – all is pretty clear. A system like this doesn’t exist to date. If we missed something, please, point it out for us. There was a system Stack Exchange founded by Joel Spolsky and Jeff Atwood, which was built on the success of the site Stack Overflow. The authors proposed a subscription-based web service, allowing creating Answers portals for your product or community. In April they changed their mind about the future of their product and said that now accounts could be created for free, but the activity had to meet a certain benchmark, defined by the owners, for a portal to launch. If you have a small service with not a lot of customers, 1000 for example, you’re out of luck, after some time your account will be closed. Maybe the topic of managing risks posed to web services and apps users as well as the ways of reducing it for the developers of such services deserves a separate post (write to us, if our thought interest you on this topic), but this event made us think about the use of web services.

In any case, because there is no such product or service called ‘Answers’, we have been faced with a choice:
- launch without Answers for now (Moscow was not built in a day).
- look at what 37signals made for themselves and make ourselves a system like that.
- create a dedicated sub-forum called ‘Answers’ on our forum platform as a patchy solution. Each new topic will be linked to a question.

Our conversations revolved around options two and three, because option one means not even trying to make it work. Developing ‘Answers’ means holding back the release of our product, which we have been making for over 2 years. Using a forum means to compromise as it is not an ideal platform to host Answers. But the arguments are still raging on. The most convincing argument is checking our system viability as quickly as possible. Having a perfect ‘Answers’ system is not critical for that. If our product is viable then we can make up a quick ‘Answers’ system and make it perfectly suit our need.

But speaking of Ideas, we have a completely different situation here. There are no free systems which you can install, but there is a load of web services. There are four possible options:
- launch without Ideas for now (Sydney was not built in a day too)
- use a web service.
- use a sub-forum called ‘Ideas’ on a forum platform of our choice
- create our own system

The thought of not using ‘Ideas’ didn’t excite us. Creating your own system would be interesting, but again we would have to postpone releasing our product. What’s left is either using a forum or a web service. So let’s talk about the web services.

Let’s use UserVoice as example. An option to allow our users to use it with a single sign on will cost us 89USD a month, or 1068USD a year. Besides their analytics is the application analytics. It answers the questions about the activity of users in the application as judged by – “At this time, voter trending, average vote counts, idea scores & activity scores are offered – in the future we intend to build these tools out to display richer data sets for power users”. It’s a good feature, but in the current phase of our product development we will be able to see its popularity with a naked eye. That kind of analytics has always been in internet forums, customer support systems, and is not related to the web analytics we are talking about. So for the money we pay, we still will not be able to combine the analytics of our product and other systems with the analytics of Ideas.

“With higher service levels we allow APIs to pull data out. In theory, you could also pull data out to generate your own metrics, or simply send the information to a CRM package.” – this is a little more interesting but … it will cost us 289USD a month or 3468USD a year. With this subscription plan, it may be possible to retrieve the data we are looking for, but you have to be ready to open your purse, and it would only do you good if you were certain if the data format you will get can tie with the data format in the forums, site and blog. Realistically you will most likely have to adjust to their format. Don’t know about you but my mood just go worse at the thought of what I will have to do. Why not let your customers generate their own custom logs in Apache and allow them to be downloaded by FTP? Maybe the people at UserVoice.com haven’t thought about this, but we await no gifts from chance. So basically you have a choice of a good service without analytics for 1089USD a year, or a good service for 3468USD a year which you can try to integrate with your analytics, spending lots of time and effort. Last option is a home-made version. It will cost nothing and be located in the midst of your forum with full integration of analytics (we have to do it for our site, blog, forum and product anyway).

We haven’t mentioned such risks of using web services like:
- change of cost or scope of services
- complete change of the product (as it happened with Stack Exchange)
- inability to add new, needed functionality later
- bankruptcy or loss of interest by the provider
- many companies on the market funded by VC, and as a result will be pushed to taking over a large segment of the market, which quite often leads to tragic results for the business.
- web service provider could be acquired by a larger company, which could completely change, or even shut down the service to which you and your customers have grown so accustomed to.

We made our choice – Ideas and Answers inside the forum, and you, if you were on our shoes, make your own which we are sure will be the right one too.

We are ready now to talk about web analytics.

Own web analytics for startups – Part IPart II

Print this post | Home

Comments are closed.