One of the questions I get asked the most revolve around how to create an analytics strategy. Most companies are already thinking of their marketing and development strategy but few tend to think about how to manage and grow their analytics tools.
The amount of data that you collect will grow over time and will come from an ever-increasing number of tools. Marketing, sales, recruitment, development and many other tools now have some kind of built in analytics component and trying to bring all of these analytics tools together is tricky.
In this post, I will walk you through how to develop your initial analytics strategy. I will also help you start thinking about how this stack will change over time and how you can prepare yourself for these changes.
Two Fundamental Principles
Before we jump into the actual analytics tools, let’s establish some ground rules for how to think about our analytics stack.
Principle 1: Start Small and Iterate Quickly
This one may seem obvious but I constantly see startups trying out all kinds of fancy analytics tools even though they don’t really need them. You don’t want to introduce complexity into your stack when it’s not needed.
For example, if you aren’t spending a lot of time or money on marketing then you don’t need 5 different tools in this area. One solid tool like Google Analytics can be enough until your needs get more complex.
Another common error is startups trying to use A/B testing tools even though they don’t have enough traffic. You’re simply wasting your time and making your data analysis more complex than its needed.
“In the world of analytics and data, [Lean Startup] means picking a single metric that’s incredibly important for the step you’re currently working through in your startup”. – Lean Analytics book
Doing some planning in terms of what KPIs you actually need will be crucial. These KPIs will be tied to your current business goals. If you’re trying to raise a Series A round then you need figure out which KPIs you need to present to investors that will help you close that round.
Avish Kaushik (from Google) has a great article on how to plan your digital marketing strategy and what KPIs you will need. I also have a short video covering how to create tracking templates for your product using tools like Mixpanel.
It’s much better to start with less and slowly add more metrics than to start with too many metrics and simply get overwhelmed with all your data. Mixpanel has a great story on a startup who actually decreased the number of metrics they were tracking to focus on what matters.
Principle 2: Own Your Data and Avoid Vendor Lock In
Your analytics stack will slowly grow to a few tools which means your data will eventually be stuck in silos. This won’t be problem until you need to start looking at your entire company, starting from sales/marketing to customer service and revenue.
In the short-term, you don’t need to worry about how this will happen. Instead, simply focus on how you can keep some kind of control over your data. A great way to avoid vendor lock in is to use an analytics wrapper such as what Segment offer. You send your events to Segment and they send them to different analytics tools.
This means you can easily change analytics tools and you could eventually even just send your data to a database like Postgres or Amazon Redshift. The goal here is to abstract your tool integrations as much as possible to make transitions away from them easier.
In terms of historical data, most tools offer APIs which let you take your historical data to other tools though this can get complicated pretty easily. I’ll cover some analytics tools below that focus on helping take your data from tools like Mixpanel to your own database.
Understanding the Different Categories of Analytics Tools
Let’s start off by organizing all the potential analytics tools into several categories. Some categories will include tools that aren’t 100% dedicated to analytics but contain key metrics that you will need to analyze how you’re performing in a specific area.
1. Marketing Tools: This category include any tools that are meant to help you optimize your marketing or advertising efforts. Examples include Google Analytics, CrazyEgg (heatmaps), Optimizely (A/B Testing), Kochava (Mobile Attribution).
2. Product Insights: This category include any tools that help you understand and improve your product by showing you what your users are doing. Challenges like product engagement, retention, and onboarding usually will require one of these tools. Examples include Mixpanel, Kissmetrics, Amplitude.
Related: We have created a free video course that will teach you everything you need to know about Amplitude. You can view the Ultimate Guide on Amplitude Analytics here.
3. Customer Support: This category includes any tools that will help you communicate with your customers and offer support. Some tools are meant to be used as part of the sales team (Olark) and some are meant to help with customer support (Zendesk). Examples include Olark, Zendesk, Intercom.
4. Diagnostic Tools: This category includes tools that will help you improve technical or usability issues with your product. You will see tools that record what users are doing, what they are clicking on and even what errors are slowing users down. Examples include UserTesting, AppSee (mobile), CrazyEgg.
5. Reporting Tools: This category includes tools that help you create dashboards and summarize the data from different tools. If you ever wanted to have a TV in your office showing the most important metrics, then you need a tool from this category. Examples include Klipfolio, Geckoboard and Excel (most versatile reporting tool).
6. Data Warehouses & Visualizations: This category includes tools that help you store and visualize data from a database. At some point, it’s much easier and cheaper to simple store all of your data in a database and just query it using SQL. Examples include Amazon Redshift, RJ Metrics, Looker, Keboola and Chartio.
7. Tag Managers: This category includes tools that help you organize your integrations and events that you are sending to your tools. Examples includes wrappers like Segment and tag managers like Google Tag Manager.
There are also other areas that you will need to think about such as sales tools (CRMs) and HR tools (employee management) but we will skip them in this post. Remember that we will start small and slowly add new tools so you don’t need to have a tool from each category on day one.
The 5 Stages of Your Analytics Strategy
Let’s start by looking at 5 stages of startup growth and the different analytics tools that you need at each stage. You may go through these stages pretty quickly so I included some helpful break points to show when you change stages.
Stage 1: All About Product/Market Fit
In this stage, you’re still trying to figure out product/market fit and you’re spending nearly all your time trying to understand your users and how to improve your product. You don’t have a lot of data so you need to rely on qualitative data at this stage.
You should also consider using something like the Net Promoter Score which is a tool that can help you measure customer satisfaction. Here’s a post talking more about how to implement it and the benefits.
- Marketing Tools: Google Analytics to give you some basic numbers around how your website is performing. Make sure to use the Goals functionality within Google Analytics to track key actions such as signing up for a waitlist or for your product.
- Diagnosis Tools: Chat tool such as Olark to start conversations directly with your users or potential users.
- Product Insights: Mixpanel/Amplitude and track 1-2 key actions within your product. You won’t have a lot of data so simply focus on a trendline (up and to the right is the goal).
Break point to next stage: You’re starting to spend significant money on marketing/advertising and you have enough regular usage within your product that you can start to make decisions using data.
Stage 2: Spending Significant Money on Marketing
In this stage, you’re starting to move past product/market fit and you might be heading down a funding round or simply looking to grow. As you start to spend significant money on marketing, you can start to deploy analytics tools to help you optimize your paid traffic.
Most of your data will live within silos which shouldn’t be a huge issue but if you need to do any complex analysis of multiple tools, Excel will be your friend. It will be tedious but it will help you avoid adding a reporting tool until you actually need it.
- Marketing Tools: Google Analytics will be your main tool for measuring your marketing efforts. You can add CrazyEgg for better heatmap tracking, A/B testing once you have enough traffic and Mailchimp to collect emails. If you’re doing mobile advertising, an attribution tool like Kochava will help run campaigns on CPI (Cost Per Install).
- Product Insights: Mixpanel/Amplitude and you can add more metrics to your initial metrics from stage 1.
- Diagnosis Tools: Chat tool like Olark to keep receiving qualitative data and you can add tools like AppSee to help track down technical errors that could be affecting your conversion rates.
- Customer Support: You can consider adding something like Intercom to communicate with customers once they are inside your product or something like Zendesk to handle customer questions.
- Tag Managers: Spend some time understanding the different wrappers to set up a solid foundation on which you could new tools in the future. Make sure to figure this category out before integrating too many tools or else your transition to a wrapper will be painful.
Breakpoint to next stage: You have enough traffic to see where your funnels are breaking and you have some ideas such as drip emails and A/B testing to fix them. You also need to see data from multiple tools in one single dashboard which means you need some kind of reporting tool.
Stage 3: Growing Teams and Dashboards
In this stage, your strategy is starting to become more sophisticated and your team is growing which means you need to create dashboard that summarizes the most important metrics. The main new analytics tools during stage will be reporting tools that help you get a better handle on the complete picture.
- Marketing Tools: Google Analytics, CrazyEgg (heatmaps), Optimizely (A/B testing) and Kochava (mobile advertising attribution).
- Reporting Tools: Klipfolio or Geckoboard to help you bring data from different tools into central dashboards.
- Diagnostic Tools: Olark (chat tool), Intercom (customer support), UserTesting (usability testing) and AppSee (mobile testing).
- Improving on anything else you currently have already implemented. This could mean adding more metrics or simplifying the key metrics you actually need.
Breakpoint to next stage: Your dashboards aren’t good enough and you want to see how everything comes together all the way from your CRM (Salesforce) to customer support (Zendesk) and then to tools like Quickbooks/Stripe.
Stage 4: Moving to Data Warehouses and Towards More Complex Queries
In this stage, you realize that some of your existing tools aren’t good enough or they don’t let you run complex queries on them. You also want to see other data within your dashboards which means you need to move your data into your own database where you can easily query and manipulate it.
- You will keep using some of your existing analytics tools across all categories but you will start to consider how this data can be migrated or stored in a data warehouse.
- Data Warehouses: set up something like Amazon Redshift or even your own database (Postgres, SQL, etc) and use something like RJ Metrics to send the data from your existing tools into it. You can then visualize the data using tools like Looker and Chartio.
Breakpoint to next stage: You have a small team dedicated to managing all of your data needs and you need SLAs, uptime guarantees and flexibility.
Stage 5: SLAs and a Dedicated Team
In this stage, you’re all about flexibility. SQL or something similar is your best friend and most of your data is in a database. You might still keep a few tools from previous stages but you need to SLAs and uptime guarantees which some of these tools aren’t able to provide.
- Data Warehouses: Amazon Redshift and an enterprise level dashboard tool like Tableau.
- There’s enterprise level solutions for tools like Google Analytics Premium but you need to balance flexibility and cost. You will also need more complex tools like DMPs (Data Management Platforms.
Planning Ahead: Overlap Tool Functionality and Data Integrity
As you can see, things can get pretty complicated quickly. You will also notice that some tools overlap in functionality but you should really stick to what each tool is best at. You can usually tell what this is by reading the marketing materials for each tool e.g. their website and talking to their sales reps.
Also watch out for potential data integrity issues. Some tools will under report numbers by 5-10% due to random technical errors. Always try to validate that your tools are collecting the right data and fix any potential errors that may come up. Data needs to be valid otherwise people won’t trust it.
Keeping Other Teams in Mind
Another thing to keep in mind is the needs of different departments. You saw that some tools are designed to help usability teams while other teams are geared towards marketing teams. Keeping in mind the different data needs from each team and how they are changing will also play a role in what analytics tools you choose.
If you got any questions or tips, I would to hear it in the comments!
I find most companies are stuck with high-level metrics and they aren't able to properly understand what actually drives user growth for their web and mobile products. To do that, you need the right data and the right tools.
If this sounds like your situation, then you should download our free tracking plan (and tutorial video). This is the document that you should create before you ever implement tools like Mixpanel, Amplitude, Segment, and Intercom. Click the image below to download your own free tracking plan (and tutorial video).