Does this mean the investment has failed? First, it's necessary to determine whether the new solution is already improving work where it is being used. If a customer can submit an order faster, an employee no longer needs to re-enter data, and errors decrease, the first signs of benefit are already there. However, as long as only a small portion of orders is processed this way, the company's overall results may change only slightly.
This distinction is important both when planning and evaluating an investment. It's necessary to anticipate how customers and employees will start using the solution, what workflow changes will be required, and how progress will be measured. Then it becomes clearer when an investment still needs time to reach the planned scale, and when it's necessary to fix what's preventing the benefits from being realized.
A mismatch between expectations and results can occur even when calculating the investment's benefits
A company estimates how much it would save if customers submitted orders independently, but the calculation omits the period during which they will start working this way. The entire projected volume and maximum savings are attributed to the very first months of operation.
Such a plan doesn't help understand whether the project is falling behind. Benefits need to be evaluated based on what proportion of orders the system will be able to process, how quickly customers will start using it, and how much manual work will remain. Then it's possible to separately verify two assumptions: whether one order really requires less work and whether enough orders are already being handled the new way.
This requires returning to the work that the investment was meant to change. When creating customer self-service, the goal may be to reduce order administration. It's not enough to provide the ability to submit an order online. The customer needs to see the prices applicable to them, select the right products, receive the necessary information, and the submitted data must flow into the company's order management system.
If an employee later re-enters each order, clarifies the price, and confirms delivery by phone, a significant portion of the work remains the same. In such cases, it's necessary to check what data, integrations, or business rules are missing for an order to be processed with less human involvement.
The investment plan must clearly state which actions will no longer be needed after implementation and under what conditions they will disappear
This allows for evaluating both the expected benefits and the work that the company itself must perform. If pricing needs to be organized, product data prepared, or approval procedures changed, these tasks must have responsible persons and deadlines.
The importance of additional organizational investments is also examined by economists Erik Brynjolfsson, Daniel Rock, and Chad Syverson. Their tresearch on the productivity J-curve explains why intangible investments that complement technologies and their benefits may be reflected in productivity indicators at different times. For a specific company, this serves as a reminder that learning, data preparation, and changes in work organization are also part of the investment.
Another condition is customer benefit. While self-service helps a company reduce administration, the customer evaluates it based on their own work. What matters to them is quickly repeating a previous order, seeing stock levels, finding out the delivery date, or downloading a document without waiting for a manager's response.
If previously it was enough to send a short message, but now they need to spend a long time searching for products and filling in information, the customer has reason to stick with their usual method. Therefore, the adoption plan starts with a specific advantage that a person will experience while performing their task.
In the study by Meuter and colleagues examining self-service trial, three readiness factors are identified: motivation, ability, and understanding of what they will need to do themselves. In practice, this means the customer must have a clear benefit, they must be able to perform the task, and understand where their actions end and the supplier's responsibility begins.
Even with a solution that has a clear advantage, customers start using it at different times. The first users, called "early adopters," provide an opportunity to see how the system works in real conditions. Their experience shows which actions are already convenient, what information is missing, and where help is needed.
The 2024 McKinsey B2B Pulse survey included nearly 4,000 B2B decision-makers. 44% of respondents were assigned to the group that relies more on relationships and traditional channels, while 20% to the group actively using new technologies. These proportions cannot be directly transferred to a specific self-service solution, but when planning adoption, customer differences must be considered.
The success of early adopters does not yet confirm that other customers' needs are the same
They may order more frequently, have better product knowledge, or have a simpler purchasing process. Before expanding adoption, it's necessary to understand whether the same functions and workflow suit a broader customer group.
For one customer, a notification about a new feature will be enough. Another will need to be shown how to place their first order. A third will only appreciate self-service when they find the information they specifically need. Therefore, after launch, it's worth talking to both active users and those who are still choosing the previous method.
A single general announcement about the system launch may not be enough here. For a customer who is just preparing a new order, showing them how to repeat previous purchases is concrete help. It must also be clear to the employee when to suggest self-service, how to help for the first time, and where to pass on an observed obstacle.
The pace of adoption is also determined by the business cycle itself. A customer who orders daily has many opportunities to try self-service within a month. Someone who purchases once per quarter may not have a single opportunity during the same period. When evaluating only the time since system launch, this difference goes unnoticed.
This is well illustrated by a sample calculation. A company has 1,000 customers. In a month, 200 of them submit one order each that can be completed in self-service, and 100 choose this method. Self-service is used by 10% of the entire customer base, but it accounts for 50% of that month's orders suitable for self-service. Both metrics are correct. The first shows the share of the customer base reached, the second shows the channel choice among those who had a need to order.
Additionally, it's necessary to know what share of all company orders can actually be completed in self-service. High usage in a small group of supported operations does not yet guarantee a significant impact on the entire business. When deciding on further expansion, both the improvement per order and the scope of operations to which it could be applied need to be evaluated.
First trial, repeated use, and impact on work are three separate things
When evaluated together, you can see where additional actions are needed:
First trial: did the customer successfully complete the order, where did they stop, and how much help did they need?
Repeat use: did the customer choose self-service again the next time they needed to place an order? Only those who already had another opportunity to use it should be evaluated.
Impact on work: after a successful order, did employees have fewer adjustments, rewrites, and other tasks left?
The number of logins or created accounts does not answer these questions. The system may attract interest, but may not yet have become the usual way of working.
The next step is to evaluate the work within the company. How much time do employees spend on one order? How much data still needs to be clarified? Are errors and repeat inquiries decreasing? By recording the initial situation before launch, similar orders can be compared later to verify whether the expected improvement has actually occurred.
If the goal is sales growth, it is also important to check the initial situation. Customers who chose self-service may have been buying more even before its launch. Their larger orders alone do not prove the system's impact: previous behavior, seasonality, and other changes happening at the same time need to be evaluated.
Once a single order is already being processed more efficiently, the significance of increasing usage volume becomes clear. The same improvement begins to cover an increasingly larger part of operations. However, when evaluating the overall result, system maintenance, user support, and the costs of remaining service channels must also be included.
Initially, the team may simultaneously serve through old channels, learn to work with the system, and help customers who are testing it. This additional workload must be anticipated in the plan. Later, it is worth monitoring whether the need for help per order is decreasing. If each repeat purchase still requires the same amount of help, the reason must be identified.
Time saved also does not yet mean costs reduced by the same amount. It primarily frees up the team's capacity. Economic benefit arises when the same team serves more orders, overtime decreases, a planned additional employee is no longer needed, or more time is devoted to sales. How this time will be used is worth planning when evaluating the investment.
Success should not automatically be equated with moving all customers to self-service. The aforementioned McKinsey study shows the importance of multiple service channels in B2B purchasing. A customer can place regular orders themselves, while discussing complex configurations with a specialist. In such cases, value arises from better distributing the team's time.
The decision to continue, improve, or re-evaluate an investment must be based on specific signals
A low number of users alone doesn't reveal where the problem lies. What we already know about customer needs, their experience, and remaining employee work helps guide the next action.
What we see | What's worth checking and doing |
|---|---|
Customers didn't yet have a need suitable for self-service | Evaluate based on the order cycle and prepare for the next usage opportunity. |
Customers who had the need didn't know about self-service | Check whether the information reaches the person handling the order; incorporate the presentation into service. |
Orders are successful, customers return, manual work is decreasing | Expand usage and check whether the benefit repeats in other customer groups. |
Started tasks consistently break at the same point | Monitor task completion and remove the specific obstacle. |
Self-service is used, but data is still rewritten and refined | Review integrations, data quality, and approval procedures. |
A single order takes less time, but overall costs remain unchanged | Assess usage volume, additional costs, and allocation of freed-up time. |
If even successfully completed tasks do not deliver the expected benefits, or the realistically achievable volume is too small to justify the investment, it's necessary to revisit the initial assumptions. Usage growth must lead to a specific business objective.
Therefore, even before building the system, it's worth agreeing on who will monitor usage after launch, talk to clients, make decisions about changes, and assess their impact. Time and budget must remain for the team to do this work. Review timing should match the actual work cycle: daily processes and infrequently recurring orders will require different monitoring periods.
For each such review, it's necessary to define what change is expected and what action should drive it. If we helped a new customer group complete their first order, next time we check their return rate. If we eliminated manual data transfer, we assess employee time and errors. This way, you can verify whether the changes made deliver the intended impact.
Time allocated to an IT solution must have a clear objective. When metrics can substantiate which process the solution has already improved and how this change is growing, management can objectively assess whether the investment is approaching the intended result.