Data Science for Business Leaders: 17 early stage tips for getting started with Analytics

When you're busy building a company while piecing together a value proposition, collecting and splicing data can seem premature and non-critical. Analytics becomes easy to put off. But before long, there comes a point where things start getting complex, and you need to understand your users a lot better, but you have with you too much unusable data since you captured it incorrectly, or didn’t bother to collect it at all. To avoid this scenario, what’s the absolute least you can get away with? We get it. You're starting out; each day brings another challenge, data shouldn’t be one of them, yet. So we need answers as cheaply as possible. Let’s explore some.

• At the founding stage, with barely ten employees, you have no resources and no time. At this point, you're close enough to the details of your business that you can still rely mostly on instinctual decision-making. All you need to measure is your product. If you’re on top of you product metrics, you can iterate faster in this critical phase. Everything else takes a back seat. Start with Google Analytics (GA).
• Also, you need real event tracking if you build software of any type. It is worth spending a bit of time doing extra research on what tools are out there. You’ll find that this space shifts a lot and even people with tons of experience will be surprised by a new tool that’s popped out. If you’re not technical, you might need an engineer to help you out with GA and event tracking.
• As a very early-stage company, the little bit of Analytics work that you do should be done with an eye to the future. Track events in a system that makes it easy for you to get your data out, as you will want to own your data at some point.
• Do not let someone sell you a data warehouse, a BI (Business Intelligence) platform, a big consulting project or anything of that sort yet. Stay focused. There is an ongoing cost when you commit to analytics. Business logic changes, data changes. It's hard to put the project on pause once you start down this road. Wait to make this investment until later.
• When you start to grow your team a bit, they’ll start needing data to do their jobs well. Everyone on your team benefits from easy access to data. Make analytics accessible to everyone in your company, not just the tech guys. This is how you make analytics a priority. Everyone should be able to look at that data and make sense of it. They may not all be data experts, so you need to make sure they’re doing the basic things right.
• You’ve probably hired a marketing person. Make sure they own Google Analytics (GA). They should be held accountable for ensuring the data is clean. Your marketing person may not be a GA person. It does not matter. If they’re motivated and smart, they can learn and figure it out. There is enough information on the web about GA. Find somebody else to do the job if they can’t figure it out.

• Build the reports in the report builder and don’t export data to Excel. We agree it's a painful process, but this will save you a lot of time in the coming months. Pick KPIs which you can easily measure within the interface.
• It’s still too early for a data warehouse and for SQL-based analytics. You need to spend all your time building, not analyzing. To run your business for now use the built-in reporting capabilities of the various SaaS products.
• Once you’ve raised your A-round and have some twenty plus employees, you start to have new options. So before you rush into any decisions, remind yourself that analytics tech is getting better, fast. Previously this type of infrastructure was reserved for much larger companies. Now you do have access to that tech at a bargain rate. This is the most critical and hardest phase, a struggle if you get it wrong, but promising when you do it right.
• Early stage companies tend to focus on ‘Build, build, build' single-mindedly. And since startups are in such a hurry most of the time, they don't take the time to understand engagement truly. As in how the product is being used, why your users are coming back and what parts of the product are working. Engagement paints a crucial picture of how users are responding to your product. There’s no way to know without looking at the data. You can no longer afford to act like an early stage company anymore. Set up your data infrastructure. This means choosing a data warehouse, an ETL (Extract, Transform and Load) tool, and a BI (Business Intelligence) tool. There are enough products in this space. Dig around.
• Bring a competent analytics lead on board. You need to find a person who can provide value from day one, someone who can also hire the team around them as the company grows. This person is hard to find, invest the time to find them. Hire someone who can think about your business and your data strategically. They’re going to be the most critical piece of your analytics puzzle for years to come. You'll save everyone a lot of time if you find someone who has an interest in analytics and task them with figuring out the most effective way to log data, to begin with.
• Can you make use of a consultant yet? Give it some thought. While it’s great that you have an analytics lead on board, he or she isn’t going to have the experience required to solve all the various analytics issues you’ll have to deal with or the expertise necessary to piece together all of the components of your tech stack. Mistakes made at this stage will cost you both time and money, so make sure your foundation is rock solid. To ensure this, today more startups have chosen to work with consultants who help them set up and then build a team around that infrastructure.
• Unless machine learning is a core part of your product, don’t hire a data scientist yet. You need a generalist, not a specialist to build your analytics team.
• Do not build your own ETL pipelines. Too many early-stage companies choose to build their data pipelines using open source software and lots of custom code. Most underestimate the effort required to build a stable system, and the ongoing maintenance necessary that system will demand.
• Do not try to get away with using more traditional databases like Postgres as your data warehouse. It’s not all that cheaper and later when you max out it’ll be a real time suck to switch. Postgres does not scale as well as a data warehouse.

• Don't be afraid of volume. People need to get over their fear of having too much data. Down the line, you're going to wish you had it. Running out of space shouldn’t be a concern at this point. Startups usually don't deal with big data, unless you turn on the Twitter firehose or you're analyzing millions of transactions per second. You can always use Hadoop if that’s the case. Even so, not many startups have this volume of data to contend with, if you do, consider yourself lucky.
• Too many in your team are still flying blind. Too many companies think they can throw data into Kissmetrics, Mixpanel or Google Analytics and they're set. But they don't really put enough thought into who on their team needs access to insights. Make a point of asking people to back their claims and suggestions with data. Everyone at your company should be making data-informed decisions all the time. Else, your engineering team will build blind while praying for the best, never really figuring out why things fail or succeed. Say you’re looking to acquire more users. You go guns blazing and implement all the proven viral app methods and pull in your existing users' in too. All of a sudden, you notice a surge in new users, which is great, but if you don’t measure the important things, you might not see how the new feature is impacting present users. Maybe people sign up because they are getting repeatedly spammed to connect their other networks to your product, and they eventually end up dropping out. Even when it's spiking churn, your tech team might see this feature as a success. Also in the process, you might have gained users who weren't worth the users you lost. These elementary mistakes happen too often. If you make your analytics infrastructure intuitive and easy for all to access and leverage to answer the questions that are most critical to their jobs, these mistakes can easily be avoided.