A data-driven framework for SaaS Packaging and Pricing
With this article, I address SaaS enthusiasts, executives and founders who look for a comprehensive and data-driven approach how to improve their packaging and pricing.
Similar to a lot of early-stage companies, Personio built its initial packaging and pricing (P&P) logic based on an educated guess. Not following a structured and analytical process. Thus, the first P&P version had a lot of flaws. There was no distinct differentiation between plans. More than 80% of customers were in one out of three plans. Customers who did not want to use all features had hardly any flexibility. Large, price-insensitive customers perceived prices as cheap. Small, more price-sensitive customers perceived it as too expensive. There was no clear expansion path. Just to name a few. On a positive note, timing could not have been better. Customers already used the product and generated a lot of data that helped to discover where value was generated.
To cut the flaws, we developed and rolled out a new P&P logic over the last two years. We learned that P&P can be a time-consuming, complex and inaccurate science. That is why many companies hesitate to touch it, leaving potential gains on the road (40% of pricing changes increased MRR more than 25%). I’d like to share our approach and provide a clear instruction for SaaS companies who plan to revisit their P&P. In the following two blog posts, I will address
- Packaging — How to develop a data-driven packaging for your SaaS product
- Pricing — How to create a powerful pricing structure for your SaaS packaging
How to develop a data-driven packaging for your SaaS product
In the first post, I focus on how to take a product apart, analyse the properties and dependencies of its pieces and assemble them to a meaningful packaging. Independent from any pricing. Eventually, customers should easily understand the packaging and perceive it as fair. For example, they should perceive the majority of features in the plan they pay for as valuable. Also, the packaging should support your go-to-market approach, for example if you are following a land-and-expand strategy. To reach that, you should make yourself familiar with different packaging components and start with a structured and detailed documentation of your product. Let’s begin with the theory.
🧩 Packaging Components
There are three common components in every packaging, which you can combine and use as building blocks: Value Metrics, Plans and Add-ons.
📈 Value Metrics
A value metric is a single measure that correlates strongly with the value that a product creates for its customers. Products with at least one value metric grow 10 to 25% stronger, especially because of higher organic expansion revenue.
You should look for metrics that buyers perceive as a good proxy for the value you are creating with your product. Otherwise customers might have trouble understanding the need for an organic upgrade. Also, value metrics should naturally move over time and be one-directional. If you use more than one value metric, make sure those are uncorrelated. For example at Personio, we use the number of employees (proxy for company size) and the number of job openings (proxy for company growth). Those two metrics hardly depend on each other. Fast growing companies with a lot of job openings and thus a rapidly increasing number of employees are particularly interesting in the Personio model. As another example, Zapier uses the number of defined Zaps (= workflow automations) and how often each Zap is used.
There are different ways to use value metrics in your packaging:
Continuous Value Metric: Single metric that measures seats, company size, usage or value-add per unit. For example, Atlassian prices seats and Zoom prices meeting hosts. New Relic prices its Log feature by daily GB for different log retention periods.
Tiered Value Metric: Single metric that measures seats, company size, usage or value-add per grouped unit. For example the GB of file storage a customer uses. Tiering value metrics can be helpful to reduce administrative overhead like billing and dunning in case the underlying metric changes a lot. For example, Mandrill prices the volume of outbound emails in blocks of several thousand units.
Value Metric Barrier: Limits the usage of a feature at a certain threshold and thus a way of incorporating value metrics in plans or add-ons. For example, Slack sets value metric barriers for GB of file storage, number of integrations, and number of available most recent messages.
📃 Plans
Plans are bundles of product verticals, modules or features. Plans should include a well-balanced mix of must-have and nice-to-have features. Usually there are 2–4 plans, but companies with a broad product portfolio — for example Zendesk — even build plans for each of their products. There are two main plan types: Either strict hierarchy or tailored to different segments.
Hierarchical Plans: Probably the most widely used option. It helps to win customers from all ends of the spectrum concerning requirements and willingness to pay. Also, it acts as guidance for a clear expansion path. Usually as options in a good/better/best packaging model. Additionally to paid plans, there is usually a free test version. Less complex products such as Slack offer a limited free version and bet on high conversion and organic expansion. More complex products like Intercom offer a free limited trial and after that lead into sales conversations. Personio also uses hierarchical plans. There are three plans tailored to requirements of different HR organisations. Additional, Personio offers the medium plan for a heavily discounted price for early-stage startups and as lifetime offering for NGOs.
Segment-Tailored Plans: This option works well for companies creating a platform with general functionality that works for several use cases. For example Hubspot offers similar functionality in tailored plans for target groups in Marketing, Sales and Service. For each tailored plan there are again three hierarchical plans.
➕ Add-ons
Add-ons are a good option to offer your customers more flexibility. Also, it can be a great source for additional expansion. Most importantly, you need to find a good balance between flexibility and complexity. Too many add-ons lead to a too complex packaging that might scare away potential customers. Add-ons can consist of single or bundled items:
Feature(s): Usually add-ons include features that only a small group of customers use, but this small group derives a lot of value from them. For example, Zendesk offers conditional ticket field in their Productivity add-on. Apart from that add-on features are rather nice-to-have than essentials.
Service(s): Remember that SaaS stands for Software as a Service. Don’t forget that the service experience around your product matters.. a lot! You should always offer a paramount service to every customer as standard. On top, you can offer premium services as add-ons to customers who are willing to pay for it, for example a dedicated Customer Success Manager. However, remember you are a software provider, not a service agency. Thus you should develop a SaaS-competitive gross margin by ensuring that services (with low margin) remain add-ons and don’t become too relevant revenue sources.
Value Metric Barrier(s): As mentioned before, value metrics can be a great option for an add-on. Hubspot’s “Reporting-Add-On” increases the limit of dashboards and reports. Other value metrics with barriers in add-ons could for example be API calls, document storage, emails etc.
Mixed: Of course, you can also mix up the above mentioned options as add-on. In order to help customers understand the additional value, think about themes for your add-ons. Personio for example offers a “Customization Plus” add-on including unlimited custom employee attributes and custom CSS for career pages.
There are many ways to package your product. A good place to start from is picking one value metric and at least two plans. But, how do you find good feature candidates for your packaging components?
🧬 Product Anatomy
Think of your product as an organism with a strict hierarchy of elements such as organs, tissue, molecules and atoms. To get an overview of your product’s hierarchy you should start with disassembling it into its pieces: product verticals, modules, features, sub-features, etc.
Be as detailed as you need to be in your documentation. List everything in a central Google Sheet which you will later use for further analysis. Name the first tab Product Anatomy.
Each component of the product has different properties. For example how users interact with it, how users perceive its value and whether it depends on other features. The structured documentation of these information later helps to identify feature candidates for different packaging components. Thus, you should accurately document dependencies, perceived value, and usage metrics.
♻️ Dependencies
On which other parts of the product do single features rely? Features that don’t rely on other features can be part of the basic plan. Features that rely on other features might be good candidates for better plans or even add-ons.
Add another column to the Google Sheet and list the dependencies between different features. For example, at Personio a lot of features rely on employee attributes.Document templates utilise employee attributes as variables to automatically generate work contracts. Thus, employee attributes is a basic feature, document templates a potential candidate for an advanced plan or add-on.
💎 Perceived Value of Leads and Customers
You should learn about value by both, taking your organisation’s and your customer’s perception into consideration.
Organization: Get input from your go-to-market and customer-facing teams on how leads and customers perceive the value of different features. As they are in contact with customers on a daily basis, they usually have a good understanding of which features drive value. To better compare the qualitative impressions of different people, I recommend to undertake a structured internal survey with a well explained scale.
Customers: Talk to your customers. One more quantitative way to retrieve useful insights is to perform a MaxDiff conjoint analysis.
This method simulates buying situations that ask users to trade one feature for another. Key is to prepare a comprehensive questionnaire and to reach a representative sample size — a certain number of customers definitely helps. The more answers the better as you can also segment answers by user or account criteria to for example understand what large vs. small customers value most. A typical question-block consists of 5–7 statements. Besides statements about features, you can also ask for other things like willingness to pay. By that you might generate additional insights about customer segments which can also educate pricing decisions.
Count the responds per statement and document them in a table. Subtract the least-important-count from the most-important-count for each feature to calculate the magnitude. You can visualise the results in a bar chart as shown below.
Again, document the insights from conversations with colleagues and customers in the Google Sheet in two new columns.
🤳🏼 Measuring Usage
For each feature, find one (or more) meaningful quantitative proxies to measure whether and how frequently it is used. This could be both, interactions in the front-end like button clicks and data entries in the database.
For example, to measure usage for Personio’s digital employee file, we look at the number of custom attributes that customers create and the average fill rate of those attributes. Invest some time as this step is crucial for further analysis.So far so good! If you have thoroughly collected and documented your product in your Product Anatomy tab, you can start composing your packaging.
🛠 Packaging Composition
Based on the anatomy of the product, you can do several analyses to find good feature candidates for different packaging components.
🔎Identifying Value Metrics
I recommend being very open about this and not limiting the selection on obvious metrics like seats. There might be superior options to drive and explain organic expansion which are not so obvious. Go through your Product Anatomy and pick 10 to 20 candidates that theoretically work as value metrics. You can start with a Feature Cohort Analysis by looking for options of features that:
- grow over time, because otherwise there will never be an organic upsell and
- fluctuate little, because otherwise customers would need to regularly up- and downgrade and
- correlate with the willingness-to-pay of your customers.
Add a new tab to the Packaging Google Sheet and name it Value Metric Usage Cohorts. For each value metric candidate, retrieve the historic development per customer for the last 12 months. Get the data from existing tools (such as Pendo) or proprietary log-tables in your database. If you don’t have historic data, I highly recommend to implement proper tracking and postpone this analysis.
Set all start dates of customers to the same period to build cohorts. So, if customer A started in February 2019 and customer C started in April 2019, both of these dates would map to period 1 in the cohort analysis. Calculate the average for each month (dark blue line). In the cohort, you can verify whether a value metric candidate steadily increases over time for your overall customer base.
It might make sense to use more than one value metric. As explained above, make sure the two value metrics are not correlated. Create a new tab Customers Product Usage in your Google Sheet. In this tab, list all your customers in the first column. List all features you consider relevant from the Product Anatomy-tab in the first row. Based on the usage measures which you defined per feature, retrieve point in time usage data for each feature. Pro tip: Make sure to use aggregated non-personal data to comply with data security guidelines.
To discover uncorrelated value metrics, create a new tab with the name Feature Usage Correlation Matrix and calculate the correlation between features based on the usage data across all customers. In Google Sheets, you can employ the CORREL() function to do so. When deciding on a set of value metrics, make sure that both value metrics tell a story that will make sense to your customers.
For example, in the image to the left you can see that the number of users and the number of documents are highly positively correlated (0.81). Whereas the number of employee attribute sections and the MB document storage per user are only slightly negatively correlated (-0.09). If you are able to find a good sales story for the latter combination, this might be a good value metric set from a pure correlation point of view.
🔎Identifying Features and Modules for Plans
For the further description, I assume a decision for a good-better-best plan structure. Start with analysing the Feature Dependencies and Perceived Value which you have documented in the central Product Anatomy before.
- A large share of must-have features should move to the good plan
- Features that rely on other features and are considered nice-to-have might be good candidates for more valuable plans
- Check out Enterprise Ready for inspiration for features that belong in Best or Enterprise plans
Extend the correlation matrix by attribute such as industry or company size and test whether specific customer segments use specific features more than others. Be creative, it’s an explorative process.
🔎Identifying Features for Add-ons
For inspiration, visit pricing pages of other SaaS companies and your competitors. Research which features other companies use as add-ons. For a good overview of SaaS businesses, visit for example SaaS Mag. Combine those findings with the data from your central Google Sheet. In your product look for features (or whole modules) that
- only segments of your customers use, for example specific reports.
- help make existing functionality better without limiting core experiences, for example search.
- add completely new functionality, for example a sandbox.
If there are value metrics that did not make it as main value drivers, you can still use them as barriers in plans and add-ons. Again, employ the usage data from the Customers Product Usage-tab and build a Usage Histogram for each feature. In Google Sheets there is a histogram chart type which helps to quickly create histograms.
In the example to the left, a large number of customers are using employee custom attributes moderately. However, a small number of customers use this feature heavily and probably also derive a lot more value out of it. Therefore, you could restrict the usage to 30 employee custom attributes in plans and offer an add-on for unlimited employee custom attributes to properly price the additional value for heavy users. Of course, you could offer the same barrier as part of for example an Enterprise plan.
💰 Packaging Visualisation and Preparation for Pricing
After analysing your product and evaluating different value metrics, you should aggregate your insights and draft packaging models. Visualising a prototype pricing page helps to challenge different models with your team.
The last step to prepare for pricing considerations is to understand how your existing customers would be distributed among packaging components. To do so, create new tabs Packaging Distribution 1.0, Packaging Distribution 1.1 etc. for each packaging model you want to test in the Google Sheet.
For each model, use the data from the Customer Product Usage tab to derive how many customers would end up using which plan, add-on etc. based on their current usage. For value metrics and barriers it is easy as you can directly use the actual data. For features, you can set thresholds to which you would assume usage. For example for the feature Scheduling employee attributes in the example above, you could define “5 per month” as a threshold for usage. If Scheduling employee attributes was in the Enterprise plan, all customers with >5 scheduled employee attributes would use the Enterprise plan in your model. Having defined those thresholds for each feature, you can start playing around with e.g. value metric barriers and see how the overall distribution of customers in the packaging would change.
This final result is a clean table listing all customers and which packaging components they would use taking the above mentioned assumptions into consideration.
💭 Closing Thoughts
There is no holy grail in packaging. You know your product and your customers best and eventually besides employing data and a structured approach, also your gut feeling should count. Independent from the technical aspects described above, a good packaging enables your go-to-market and customer-facing functions to tell a comprehensible and exciting story. Thus, the creative and last part of the packaging exercise is to block half a day with a cross-functional team (e.g. Product Marketing, Design, Sales and Customer Success) and create a profound story around your packaging.
Good luck with your packaging project! Next up is pricing: How to come up with reasonable price tags for your packaging components.
