Test new bundle configurations

Do not invent bundles in a workshop. Look at what customers already buy together. If you have not done that yet, follow the pairing method in how to structure cross-sell offers.

Write the hypothesis in one sentence: "Customers who bought A will buy B if we offer them together as a single package, because B solves the problem A creates." If you cannot say why, pick a different pairing.

Test no more than three. These shapes cover most cases:

• Core plus one add-on: your main offer with the most common second item attached.
• Good, better, best: three levels of the same bundle with growing scope.
• Outcome bundle: items grouped under a result, such as "ready for your first quarter", rather than a list of parts.

Name each bundle by the problem it solves. Customers buy outcomes more easily than inventory. For how to set the tiers and avoid cannibalising your plans, see how to design upsell paths.

If you change what is in the bundle and what it costs at the same time, you cannot tell which one moved the result. Keep the price logic the same across configurations in the first round, and test the contents. Test price afterwards, using the approach in how to price expansion and growth.

Choose accounts that fit the offer: customers past onboarding, with a visible need. Split them into groups of equal size by alternating down a list sorted by account size, so each group has a similar mix.

For a base of 60 customers, that is 20 accounts per group at most. That is too small for statistics, so treat the result as direction. Back it up with conversations.

Write down, in advance, what counts as a win. Use these measures:

• Attach rate: accounts that buy divided by accounts that were offered.
• Time to decision: days from offer to answer.
• Discount used: how much you had to give to close.
• Revenue per account over 90 days.
• Retention and support load after 90 days.

Example rule: "A bundle wins if the attach rate is at least a quarter higher than the next one, and no more than a tenth of accounts need a discount." Setting this first stops you from picking the favourite after the fact.

Add a short question to every decision: "What made you choose or skip this?" For buyers, you learn what they valued. For those who decline, you learn what was missing. Put the answers in your CRM, such as HubSpot, next to the offer so you can count them later.

If you want a quick read on willingness to pay, ask two questions in a call: at what level would this feel too expensive to consider, and at what level would it feel like good value. Compare answers across accounts.

Give each offer time to reach a decision, especially in larger accounts. Review weekly, but do not change anything during the run. At the end, compare the criteria you set. Then keep the winner, drop the weakest and design a new challenger.

Record the result and the date in your expansion revenue tracking so that you can see the effect over time.

• Testing too many configurations at once.
• Offering the bundle to accounts that have no need for it.
• Judging the result on attach rate alone. A bundle that sells and then causes complaints is not a win.
• Giving discounts to one group and not another.
• Stopping after one round. The first winner is rarely the best one.

You can name the winning bundle and show why it won against criteria you set beforehand. Its attach rate holds when you offer it to the next group of customers, and those customers still renew after 90 days. You also know which configurations to retire, which is as valuable as knowing the winner.