Setting Optimal Pricing and Packaging

Why does good pricing matter?

It shapes customers’ conception of your product’s value – if you start at a very high or very low price point compared to other products, it’s a signal to customers about how much value your product offers.

Price is your biggest lever to affect your margins – people are sometimes afraid to talk about margins when it comes to pricing because they want to avoid a “cost-plus” pricing method. But in reality, price is the biggest lever you have to affect margin. Investors will benchmark your gross margin against all of your peers, it’s the first indicator they have to understand customer perception of your company and significantly affects your ability to create a valuable company in the long run.

When should you begin taking a structured approach to pricing? What are the downsides to starting too late?

Start thinking about pricing before you launch – it’s never too early to start thinking about pricing. Pricing affects a customer’s perception on day one, and if you develop an understanding of how your product should be priced, you’ll also get a better idea of how your product can deliver value to the market.

It’s very hard to go upmarket once your pricing is set – once pricing is set, you’ll have a hard (but not impossible) time going upmarket from the initial price point. Customer perception is set from the beginning. The longer you scale with suboptimal pricing, the harder it becomes to change customer perceptions.

Think about pricing to ensure product development delivers value to customers – without a pricing exercise that gives you in-depth knowledge of what your customer values, you may miss certain product features, or build some that customers ultimately don’t care about. Think about price and value while you’re still building the product or adding features.

How should you think about the way you price and your unit of value?

You can monetize in numerous different ways:

• Single purchase basis – a single upfront purchase of your product. Rely on repurchases over time to maximize revenue per customer.
• Subscription – fixed payments on a recurring basis, either a flat fee or a variable one based on number of seats, etc.
• Usage-based pricing – payment varies according to usage from customers.
• Dynamic Pricing – this can be used across any of the above pricing methods, but is important to include here as it is truly the best way to optimize profitability and capture true willingness to pay. Dynamic pricing changes pricing based on factors like demand vs. capacity, daily usage, etc.

The right method depends upon your product and customer:

• If usage is very stable → subscription-based pricing – if usage is really consistent, don’t over-complicate things; a flat subscription may be intuitive and fair.
• If usage is very variable → subscription-based pricing – unintuitively, if usage is highly variable, you might not want to tie pricing to usage if financial predictability and consistency is important to your customers.
• If usage is correlated to value → usage-based pricing – if the value customers get correlates tightly with usage, a usage-based pricing metric allows you to capture more value from your most successful customers.

Let the customer guide you to how they prefer to pay – if you ask a customer whether they’d rather pay $5 a month, or $0.50 per usage that works out to be $5/month, and your customer base prefers one over the other, listen to them. Most often, you should just follow their lead unless one pricing method clearly allows you to demonstrate or capture outsized value.

How is segmentation related to pricing?

Segmentation is one of the most important aspects of pricing – every one of your customers or potential customers is willing to pay a different price for your product based on their perceived value of your product, specific use-case, demographics, or even some obscure behavioral or attitudinal attribute. There is a global pricing elasticity curve for your product where one axis is price and the other is the total number of customers willing to buy at that price. At some point, you start losing potential customers for every $1 increase in price. But if you can break up customers into segments and price differently for each, you’ll be able to capture an even larger portion of your customer’s willingness to pay.

Segmentation finds hidden value in your customers – if you segment well, you’ll find useful insights like groups of customers that use the product differently than everyone else or that will pay a lot more for the product if it has certain features. These improve your product pricing, bundling, and positioning.

How should you think about bundling?

Bundling helps you package and charge different prices to different segments – bundling features together that matter to specific segments allows you to change pricing for that bundle to align with the value the customer receives. Bundling is one of the ways to align your product uniquely to each of your customer segments, and ultimately align the value that your product delivers to the top of their willingness to pay.

Don’t put every feature in every bundle, even if every segment might like it – simply because there is a feature that all segments express interest in does not mean that it’s the most important feature for that segment. Find the “must-have” features for your most valuable customer segments. If that feature is not “must-have” for every other segment with a lower willingness to pay, don’t include it on those bundles.

“Good, better, best” is one of the more basic but easier to implement packaging practices (suitable for companies with <$20M in revenue) – companies do this so they have one offering for SMBs, one for corporate customers, and one for larger enterprises. Earlier-stage companies might not have the time or resources to build more complex customer profiles, so they segment using basic demographics. Even though this is very basic, it is not a bad starting point – and this very basic segmentation is still very beneficial as you work towards a more intensive pricing methodology.

Tip: Add more features AND a higher seat minimum to the best offering – take the features only an enterprise buyer would want and bundle them with more seat licenses so that enterprise customers not only pay the most per license, but also buy more seats (for which they will presumably have more employees to fill).

What types of research activities should you use to collect data for pricing?

Companies should use surveys, interviews, focus groups, mystery shopping, win/loss, and 1:1 calls – a great pricing strategy requires a company to deeply understand its customers. Customer-obsessed executives realize very early on that they have pricing power because pricing and product value are intrinsically related. Listen to the customer so you can deliver to them the most possible value and price accordingly. This is impossible to do if you aren’t speaking with your customers, often.

Start pricing research pre-launch with 1:1 calls – these are largely to get the voice of the customer into your product. It could be prototyping, getting feedback, and asking potential customers if they would buy something like your prototype.

As your company scales, move into more formal 1:1 calls, surveys, and focus groups – as you grow you can start to collect feedback from customers in more standardized and formal channels. Expand from your current customers to prospective customers, to a wide universe of unknown but potential customers that might not have heard of you yet.

Test all of the important questions for a survey in 10-20 live calls first – you’ll learn a lot as you test questions to ensure you ask the right things. Oftentimes a follow-up question reveals something that changes the way you ask or analyze survey data.

Customer interactions should be an ongoing habit – it doesn’t have to be a product manager scheduling hour-long calls; you should seek out conversations and feedback from customers whenever you can. Predicting customer behavior is at the core of your business’s success. And there isn’t a single customer. There are dozens of different types of customers, with different needs, who all value things differently.

Users and buyers both matter – for software, it’s important to remember that the buyer and the user might not be the same person and you should talk to both the buyer and user to understand how their priorities might differ.

Who should you solicit feedback on pricing from? 

In order of importance, solicit feedback on pricing from:

1. Ex-customers – these are the most valuable conversations. Someone used your product, decided it wasn’t for them, and probably went to someone else who they view as a better value.

2. Current customers – they can help you understand how current users derive value from the product and how they communicate that value internally. Be aware that they’ll probably report a slightly lower willingness to pay than reality. Sometimes it is best to contract with a third party for these conversations to gain transparency from current customers.

3. Prospective customers – they have been in the pipeline and understand the product enough to have a perspective on the product’s value, but the same caution for current customers might apply to some in this group as well. It’s also not generally a good idea to survey this customer set if they are late in the pipeline as the survey or interview may disrupt the more important sales process.

4. Market – sourcing opinions from people who don’t understand your product provides limited insight. If I asked how much you would pay for the best glass of orange juice in the world, you wouldn’t be able to tell me because you’ve never tasted it. It’s even harder when trying to describe a unique piece of software. Still, this group should not be ignored, especially as you try to gather larger data sets – it simply requires more time to inform the respondent about the product up front.

5. Ex-competitor employees and customers – you have to go up against somebody else’s pricing in the market. A perspective on your competitors helps you understand them, but you’re better off spending time understanding your customers and how you can better meet the market’s needs.

Good pricing research requires respondents to understand your product – insights from surveys and interviews are only as good as a respondent’s understanding of your product. If you ask hypothetical questions about willingness to pay for an imagined product, the survey is effectively useless. It is also important to consider how you screen your respondents. Ideally, you’d like your sample to match the rough demographic of your prospective or desired customer set. For example, if you ask a group of 60-year-old men about their willingness to pay for a skateboard – your results will be skewed (unless, of course, you have a new innovative skateboard specifically targeted at 60-year old men).

One last warning: If you are having individual 1:1 conversations, don’t over-anchor on anything a single individual says. You may have an non-representative customer on the line who is more or less price sensitive than the median; or they may belong to a specific target customer segment – and you don’t want to change your pricing strategy based on the anecdotes shared in a single conversation. All the things you learned about sample sizes in high-school or college still apply here.

What questions should you consider including in your research?

Segmentation Questions

Pricing Intelligence Questions

How many data points do you need to inform your pricing (especially if you’re performing analytics like conjoint)?

If you’re not doing segmentation, you should aim for 150-300 responses – the larger the sample, the better. But you need a representative sample of buyers. If you sell to enterprise and you get 1,000 responses from SMBs, the data will be worthless – so it’s not simply about volume.

If you’re doing segmentation, you should aim for more than 200 responses from each segment – if you believe you’ll have three to four segments, you should aim for 600-800 responses at a minimum (assuming all segments are of the same size), and your data doesn’t get very safe until you’re near 1,200 responses. The larger your company gets, the more detailed your segmentation can become.

When pricing in a large enterprise and small n market, you’re better off doing 1:1 phone calls – pricing when neither you nor your market has many firms gets a lot harder. A panel or third-party facilitator can help you get responses but using your own pipeline won’t get you there. You’re probably better off using one-on-one phone calls to dive deep on questions and extract meaningful answers.

How should you use buyer personas and cluster analysis to segment customers?

Buyer personas humanize the data for the sales and marketing team – the endpoint of your research is data-backed ideal customer profiles and market segments. They arm your sales team with humanized personas to sell into, rather than opaque raw data. Personas will guide sales to ask the right questions to identify segments, and then position the product to meet the needs of the segment they’re talking to.

A buyer persona might look like:

• The title of the ideal buyer for that segment
• The firmographic traits of an ideal company in that segment
• The attitudes and behaviors that this title is likely to have
• Their spend/budget in this area
• The features they value the most

Personas help you match prospects with packages – sales uses personas to guide how they talk to prospects and position your product. Marketing and sales can leverage personas to create language and positioning that aligns expression of the product’s value with the needs of the segment.  

Cluster analysis will help you build personas – the most accessible version of cluster analysis is finding a Python Script on Google that runs a K-means algorithm across your data. K-means will run hundreds of times across your data to identify the most distinct traits that cause certain groups of respondents to comprise a discrete segment. It still takes a lot of human effort to identify what makes those customers unique and search for the cause of the clustering.

Using 3-5 segments/personas – it’s possible to have seven, but beyond that it gets unwieldy. 2 segments is an acceptable start, but is probably not robust enough to capture all of the value in your market. Identifying three to five segments will really allow you to price discriminate between segments.

Who owns the pricing exercise? Who else is involved?

CEOs need to be customer-first and care about pricing from day one – if you don’t have a CEO who cares what customers think and is in tune with the customers, it’s hard to convince them that pricing should be prioritized. A CEO may even want to run some in-depth interviews with customers or sit in on a few pricing calls.

The pricing exercise can belong to several different functions – it could be marketing, sales, a chief of staff, product, or a corporate development/BizDev employee. Even at a large company, you may not find anyone with the internal expertise to handle a pricing project, hiring an outside consultant works fine. Get consensus from the executive team.

What are the steps to executing a pricing project?

Step 1: Build a market map and sit down with engineers and product management – seek to understand where your product sits relative to the market and how the value of your offering is differentiated.

Step 2: Have conversations with customers to understand how they perceive value – these might be conversations where you’re discreetly trying to figure out pricing. Ask about the value of your product and get their feedback so you understand what they care about. Then slip in pricing questions.

Step 3: Hold more formal 1:1 interviews – you need to test out any questions you plan to ask in a survey in 10-20 conversations with customers before using them. This will help you ask smarter survey questions and gives you a chance to ask follow-up questions to dig deeper into answers.

Step 4: Roll out a survey – for it to be effective, you need to communicate the value of your product in the survey. You might include video, pictures, or 3D renderings of the product so that you’re not asking theoretical questions.

Step 5: Build your ideal customer profile and bundles – ideal customer profiles are the deliverable that you hand over to your sales and marketing team to help them understand the different customer segments so they can better close deals with them. The bundles should be targeted to meet the needs and capture the willingness to pay of your customer segments.

How should you roll out, test, and adjust pricing?

Intentionally evaluate pricing every 2-3 years – you can keep running the same surveys and make minor tweaks to pricing yearly, but customers aren’t going to accept major price changes more frequently than every 2-3 years. If you are operating in a fast-moving market, you may have to do this more often as the segments grow and change, and new ones arise.

You should constantly test your pricing – it might be as simple as taking a couple of sales calls where you double your pricing every time you go into a sales call until someone laughs in your face.

A new product or feature launch can trigger a refresh – these big launches tend to happen every 2-3 years anyways and they provide customers with a narrative reason for why pricing might be changing.

What are the most important pieces to get right?

Express your product’s value proposition effectively and ensure understanding – this part matters! I’ve gone so far as to run a survey with a group of 20, then called them afterward to check if they understood the value prop and the product. If you don’t nail the expression of your product, respondents can’t properly express the value and willingness to pay back to you.

Data informs the strategy, it doesn’t dictate the strategy – this is probably the biggest criticism of pricing studies from execs. Someone runs a pricing study and makes a presentation saying you have to do X. But there is typically a world of considerations not captured in the pricing data—data shouldn’t dictate the strategy, it should inform it.

What are common pitfalls?

Overreliance on a single data point – for example, if you do Van Westendorp and every single data point says that at $50K the product is becoming expensive but isn’t too expensive—yet you’re having success selling at $100K, it’s important to consider both pieces of information before reshaping your entire pricing strategy. It also likely indicates that your segmentation is still not robust enough.