It happens over and over and over again. A company has a successful growth spurt and is ready to ratchet up to the next level. They are sitting around a table trying to decide whether to add more offerings, enter adjacent markets, raise their prices, etc, etc, etc. Very quickly they realize it would be ideal if they had analytics and metrics on A, B and C to help make the best informed decision. But they don’t have A, B and C because they didn’t capture the data during the earlier days and now it’s too late to start capturing the information for the needed near-term decision.
When to Start
This dilemma happens at all stages of company evolution. So why not instrument everything for data collection from the start? Seriously, data storage is unbelievably cheap so that’s not the inhibitor. The hardest part is deciding what information/metrics to collect and archive. My answer in the early days of generating revenue is EVERYTHING. Now you just have to figure out what “everything” means. Let’s explore further.
- Usage information
Which modules and features are being used by each user, how often, how much (quantity), etc. This is especially useful for your highest-value features that sometimes go unused. What are the attributes of the user that uses those features? Info from freemium and free trial users is also really valuable when trying to determine the behaviors that lead to conversion (to a paid offering) - Time-based data
How long before customers start doing A, B and C - Demographic data
Geography, company size, industry, job title - Customer Lifecycle data
Especially things related to customer acquisition (sales and marketing), service, support and hosting
The list obviously could go on and on but these are some examples to get you thinking. Again, capture it all and you’ll be amazed how frequently you’ll say in a planning meeting something like “wow, I’m glad we actually thought to capture this information”.
Start with Business Questions
As part of this process, try to identify some business-related questions you already have that could be answered using historical data. Then, imagine yourself pitching to a Series A VC for a round of funding. What questions are they likely to ask to better understand your business and to gauge your level of understanding of the business? If you don’t know what types of questions they’re likely to ask, either meet with such investors to find out (relationship building), talk to your advisors, read most of the articles in the fundraising section of this blog, or order my bestselling book on fundraising. Actually, do all of those things as part of this process.
Below are some examples of business-related questions that might relate benefit greatly from historical metrics:
- How do we know if a freemium or free trial user is likely to convert to a paid offering? How do we know if they’re likely to churn?
- How do we know if a paying customer is a good candidate to upgrade to a more expensive offering? How do we know if they’re likely to churn?
- What is our customer acquisition cost (CAC) and associated CAC payback? (see related article titled “Visualizing the Interactions Between CAC, Churn and LTV“)
- Should we raise our prices? Should we lower them?
- What are the attributes of our best customers? What are the attributes of our worst customers? [of course, you have to decide what “best” and “worst” means]
- If we had an extra $10K to spend on marketing, how would we spend it and why?
- If we were able to double the size of our sales team, how would we allocate them and why?
Hopefully you get the idea for how many areas of your business plan can be covered with this exercise. Once you have the list of questions, work backwards to the types of historical data that can help answer the questions.
Also remember that your initial business plan is actually a huge list of assumptions (read my related article by that exact title). One of the first things you need to prioritize is validating those assumptions. So if you follow the suggestions in the article I referenced, you’ll already have your first list of business-related questions to answer with collected data.
Capturing and Processing the Information
This information will come from your financial system, hosting platform, CRM, marketing automation platform, website analytics service, etc. But this creates an added burden of consolidating the data for analysis and correlation across data types. Check out metrics aggregating vendors like Mixpanel and KISSmetrics. Or check out Heap Analytics’ technology that makes it easy to instrument any element of a web page with the click of a button and without requiring the web developer to be involved.
Here’s a report from someone who compared these two to give you an idea about the capabilities they have: See Report Here. More recently, AirTable is getting a lot of attention as an aggregation and analysis tool. I’m not promoting one of these tools over the other and I also realize vendors come and go over time. So use the comparison report I linked to for general education on the topic of operational instrumentation.
There will be a day when you have so many business-related questions and so much collected data that you’ll add an Analyst to the team. I think they can be worth their weight in gold and should probably be added to the team sooner than most startups decide. Combine this with the dramatic impact machine learning and other AI-related disciplines can have on a business and the justification for adding this role is further increased.
Summary
I recommend making this whole topic a part of your company’s culture. Train your managers to start most of their strategy and planning meetings with business questions. Educate them on the vast amounts of data you’ve collected to help them answer those questions and run a well-oiled machine for their function. Arm them with systems and tools to make all of this really easy and saving the need to pull in the Analyst for really difficult questions.