The pervading approach to launching a startup is to do it quickly, iterate constantly and make as much noise as possible throughout the process. It’s not a bad way of doing things, and given the lower cost of startup operations, and the nature of consumer web startups in particular, and it’s completely doable. But be careful if you’re not a data hog.

Getting your startup launched as quickly as possible is fine – you need to get it in front of people to understand what’s working (and not), and get as much feedback as possible – but you should also spend a good chunk of time preparing to collect data. That means building the necessary infrastructure into your system to collect, review and analyze the data generated by users.

What Data Should You Collect?

Anything and everything. Collect as much as you possibly can, even if you’re not sure of its value upfront. Data has a sneaky way of revealing things over time – things you might not have thought of immediately. Data has a way of helping you figure out what questions to ask, because it exposes trends, and allows you to look at things with different perspectives.

Ask Basic Questions to Start

Start by asking yourself some basic questions on how you expect your application (or hope your application) will be used. There are some fairly common questions and data points that will be of value regardless of what type of application you’re building (be it B2B, B2C, etc.) For example:

• how often do people log in?
• how long do they use the system?
• what features are people using?
• when are people using the system?
• where are the users located geographically?

If you’re out of the gate with a business model and charging customers, there are a whole bunch of additional questions you can ask:

• how many people are paying?
• what are they paying for?
• what payment plan are they using (if you offer monthly, yearly, etc.)?
• how much are they spending?

Questions Beget Questions

As you start to ask questions and answer them with the data you’re collecting, it will lead to more questions. Getting into an analytical mindset of evaluating trends through data will help you uncover all sorts of interesting things. Here’s a good example from Standout Jobs…

We currently distribute job postings to a variety of job boards and job aggregators, including SimplyHired and Indeed. We also sponsor jobs on both SimplyHired and Indeed (through a pay per click model) to see how well those jobs perform in terms of generating clickthroughs and applications. And we want to compare the two of them.

A few simple and obvious questions we ask (and collect the data for) include:

• how many clickthroughs are generated from these job aggregator sites?
• how many applications are generated?
• how much is it costing us per click and per application?

What’s interesting is that we notice a higher clickthrough and application rate for new jobs that get submitted through our feeds into SimplyHired and Indeed. That makes sense, of course, because people are always looking for the freshest jobs. So that leads to the next question, “How many clickthroughs and applications do we receive for jobs over time?”

This is interesting because it can affect how we spend money on sponsoring jobs. If we see that a job receives almost no applications after it’s been in an aggregator for 2 weeks, why bother paying for it show up anymore? So that leads us to think about optimizing our spending based on the age of jobs…

That leads to a whole bunch of other questions, all of which are answered through the data we collect.

Data = Money

You can make money from data in a few ways:

1. The data itself can be valuable. People will pay for data if it helps them answer questions they need resolved. It’s really as simple as that. And entire businesses have been built on collecting data and reselling it, or selling the knowledge gained from the data.
2. The data can optimize your business. You can use the knowledge gained from data to become more efficient and innovate, which will save you money. And saving money means making money.
3. The data can lead to new business opportunities. Simply understanding what parts of your product people use can help you find ways of staying focused and making more money from it.
4. The data can drive product development. You may even discover new products worth building based on the data you’re collecting.
5. The data can drive sales. For example, we track “last login” for customers that haven’t yet published their career web sites. When we see a prospect that’s recently logged in, we get in touch to see how things are going, and very often can convert them on the spot.
6. The data can improve customer support. Fixing bugs is always frustrating when you don’t really know what a user was doing. And as much as you’d like them to tell you, they can’t always do a good job of it. If the data can help you figure out how somebody was doing something when they ran into trouble, you’ll be able to fix it faster. That’ll make your customer happy. Happy customers spend more money.

Startups need to collect data.

Incidentally, VCs love data. They understand the value behind it, and how entire businesses can be discovered, created and evolved off of collecting lots and lots of data.

Data Doesn’t Always Tell You Why

Data can tell you a lot of things, but it doesn’t always answer the question “why?” Answering “why” typically requires more analysis of what’s going on, a deeper understanding of user behaviors, some guesswork and investigation. Don’t be afraid to go to your customers and ask them “why” — often they’ll be happy to tell you. And then you can correlate user answers to what you see in the data, and make the best decisions from there.