2 of 8

Why should we care?

Why does this difference matter? For starters, we should be verbal about industry terms like this. Deflection and Resolution come up all the time for our prospects, and we need to understand each of them in depth.��More practically, we need to understand the difference from a competitive standpoint. We see competitors like Espressive, Aisera, and others talk about the importance of measuring deflection, their sky-high deflection rates of 90%+, and even compare those results to our resolution rates. In some cases, this may come from a place of innocent confusion, in others, it may be an intentional strategy to eat our lunch.

In any case, we all need to be prepared to ensure our prospects are measuring these success metrics apples to apples, and that we can be seen as the arbiters of clarity in the market.

3 of 8

The deflection experience

Deflection

Exit chat

What’s the Wifi password?

Let’s explore what it means to be deflected. Let’s say you’re a brand new employee arriving at a new office on your first day. You settle into your new desk after being told where to go by the front desk, only to realize you don’t know what the wifi password is. You begin as that neutral emoji, wondering what to do. You don’t know where your IT help desk is, so you can’t take that path, but they you remember, hey! I learned that we have a chat bot for stuff like this. You go there and type in the question, “What’s the wifi password?” There are two potential paths from here, either you get the answer and you’re pleased with yourself OR you get a knowledge article about connectivity issues, but no password, and you get mad. Stupid chatbot.

Whatever happens, you exit the chat. Because you went through the the bot, asked something, and exited, that counts as successful deflection. Doesn’t seem like the best experience though, does it?

You still don’t have the wifi password, so you take the next bottom path first, turning to the person next to you. They’re a vet of 5 years and don’t remember. You then turn to trusty google, which, as it turns out, is unable to search your intranet. Finally, you search around your office, finally stumbling across your IT department after 15 minutes. They know you answer.��There are two issues with this journey. First, even though you may have not gotten the answer you needed, it still counted as positive deflection. Second, you still had to waste your and the IT department’s time by figuring out how to ask this question, and what the answer was. Again, this is a terrible user experience.

4 of 8

The resolution experience

Exit chat

What’s the Wifi password?

Resolution

5 of 8

How they differ mathematically

All requests handled via only chat with bot

# of requests not transferred to agent

# of requests fully resolved by a chatbot

All requests handled via any support channel (e.g. chat, phone, email, etc.)

Measuring deflection ignores the employee experience side of technical support.

Measuring resolution gives accurate insights into your overall employee satisfaction.

30% - 50%

Moveworks Average

Deflection Rate

70% - 90%

Industry Average

Resolution Rate

Another way to understand deflection and resolution is to dissect the equations behind them. The industry defines deflection as the number of requests NOT transferred to an agent divided by all requests handled via chat with bot.

On the other hand, Resolution Rate measures the # of requests fully resolved by a chatbot divided by all requests handled via all support channels, including the bot, email, phone, and more

Again, I want to be clear about this. Where resolution rate correlates with employee satisfaction and bot adoption, deflection is completely unrelated. The bot could have a 5% adoption rate and show a deflection rate of 90%+. However, resolution rate would reflect that super low bot adoption rate.��Sources:

https://www.meya.ai/blog/deflection-rates-chatbots

https://blog.mavenoid.com/deflection-vs-resolution-why-chatbots-struggle-in-tech-support/

6 of 8

How they differ by definition

Moveworks uses Resolution Rate to measure the success of our IT product because we care about solving your employees requests, not just deflecting them to figure it out for themselves.

Deflection Rate

The percentage of support requests your human support team did not have to address

Disadvantages

Does not capture instances when requests are deflected but not solved
Cannot be used as a corollary to customer satisfaction with your chatbot
Easy to obtain high percentages that aren’t insightful

Resolution Rate

The percentage of support requests fully resolved by support automation

Advantages

Captures instances when requests are automatically solved from end-to-end
Can be used as a corollary to overall employee satisfaction with your chatbot
Metric is a reflection of saved time and cost

To drive this point home, let’s review the verbal definitions, disadvantages of measuring deflection, and advantages of measuring resolution.

Deflection really measures savings in the agent’s time, the percentage of support requests your human support team did not have to address. What are the disadvantages of this?��1. It misses cases where requests are deflected but not solved

2. It cannot be used to also track employee satisfaction

3. It’s very easy to obtain a high percentage that yield no insights

On the other hand, Resolution measures the percentage of support requests fully resolved using automation, saving not just the agent’s time, but the user’s time as well. The clear advantages of measuring this include:��1. It only captures the experiences that are solved end-to-end

2. It can be used to track employee satisfaction

3. This metric is a reflection of saved cost and time

Moveworks uses Resolution Rate because of these advantages – we care about creating a magical employee experience.

7 of 8

Comparative examples in the wild

Deflection Rate Examples

Resolution Rate Examples

When an employee asks for access to SFDC, automatically providing access to SFDC.

When an employee asks for access to SFDC, providing an email address for them to reach out to.

When an employee asks about their parental leave and are sent a knowledge article that may be correct or not.

When an employee asks for parental leave, are sent the correct snippet of text, and it is not labelled “unhelpful.”

When an employee is trying to create a distribution list but becomes frustrated, exits the chat and asks a colleague instead.

When an employee tries to create a distribution list and is able to automatically manipulate an email list through their IDAM end-to-end

Finally, let’s walk through some examples, besides the “where’s the wifi journey”.

Let’s say you need access to SFDC. In a deflection example, you’d be sent an email address to contact, and that would be considered successful deflection. This example is exactly what one of our competitors does. But that’s not resolution. Resolution in this case, would only be successful if you were actually given access to SFDC.

Maybe you need to ask for parental leave? Just being sent a knowledge article counts as deflection. However, with Moveworks, to obtain successful resolution, the employee would need to not only receive the correct snippet of text, but also not label the response as “unhelpful”.��Finally, let’s say you’re trying to add someone to a distribution list, and you try it, become frustrated because it doesn’t work, and exit the chat. That’s considered successful deflection. Again, for this to be a case of successful resolution, you would need the bot to take action on your request and add the new person to that distribution list.

8 of 8

Track a collection of metrics – not just one

Resolution Rate

The primary goal for most virtual agents is to solve as many support issues as possible, automatically, with no intervention from a support team.

Time to Value

Consider how long it will take for your system to deliver meaningful results for your employees, your support teams, and your company as a whole.

Time to Resolution

Resource Utilization

Bot Preference

Issue Prevention

Satisfaction Score

Your virtual agent should accelerate support, so employees never have to wait for help. This is different from Mean Time to Resolution (MTTR)

One of the most important metrics to think about is how employees are using – or not using – the knowledge base articles and forms you create.

Bot preference measures what percentage of employees use the virtual agent as their primary way to get support

Issue prevention tracks the percentage of issues your virtual agent can prevent by sending proactive messages and targeted communications.

The Moveworks bot was designed to positively transform the employee experience - consider this to measure whether employees like the solution.

Metric

How It’s Measured

Example

40%

8 Weeks

24 hours

Varies

85%

15%

90%

Now, as I alluded to at the beginning of this discussion, deflection and resolution are not the end-all, be-all metrics for ITSM. At Moveworks, while we care about Resolution Rate, we also care about 6 other key measures.

Time to Value is vital for our customers to achieve quick results and obtain a return on investment on their purchase of Moveworks. Many competitors, particularly toolkits, struggle against this particular differentiator.��Time to resolution measures how quickly a request is resolved. Ideally, adopting a chatbot like Moveworks saves time for both agents and employees and cost for the company overall.��Resource utilization measures how much employees are looking at, learning from, and using all the knowledge that has been built out by the IT team over days, months, or even years. It proves out the value of those resources and ensures they’re being utilized.

Bot preference measures how often employees go to the bot for support first, as opposed to other channels like email, phone, or in-person. This again indicates how successful the chatbot program is for the IT help desk.

Issue prevention measures how often the virtual agent is able to avoid support requests by being proactive, sending a targeted message or nudge about specific, relevant topics.��Finally, satisfaction score is pretty self-explanatory – it measures how employees feel about the solution.

At Moveworks, we closely track and care about all these metrics – not just resolution, and certainly not deflection.

I hope that this presentation helped delineate the difference between these two often talked about metrics. It’s complex and our competitors often purposely muddy the waters to make their solutions look better. Let’s aim to be thought leaders and educators in the space and dispel with their FUD.

Good luck out there and thanks for listening!

1 of 8

2 of 8

3 of 8

4 of 8

5 of 8

6 of 8

7 of 8

8 of 8