Measuring P2P support

  • 19 December 2017
  • 4 replies
  • 203 views

Userlevel 7
Badge +4
How do you measure peer to peer support and how important is this metric to you?

Here are a few reasons why I think the customer helps customer % metric is skewed and unreliable.

It could be low, because:
- Customer helps customer % is often defined by the effectiveness of your moderator team. If moderators are more speedy and effective than usual for whatever reason, chances are peers will provide less marked answers relatively.
- Summarizing and quoting the tips and ideas of others and marking it as an answer can help a topic being more useful and easier to read. It's not counted as peer to peer support, while really the community helped out.

It could be high, because:
- Some customers mark their own response saying 'Customer service fixed it for me' as an answer. The support clearly doesn't count as coming from peers, but it's counted as such in customer helps customer %.
- Other customers are good at redirecting other customers quickly. I'd rather have 100 more visits to service articles than 100 redirections by other customers to these service articles marked as an answer in a topic. These redirections count as P2P support, but really they're really due to bad search performance.

Also:
- The act of marking an answer isn't taken into account in this metric. In true peer to peer support, moderators aren't involved at all imho.

There really isn't anything better than customer helps customer %, but I'd like more in depth analysis of Likes to complement this metric.

4 replies

Userlevel 7
Badge +1
Hi Jurgen!

Thanks for sharing your opinion on this, I understand why you may see it that way.

From my personal experience however, some of the factors you mentioned should not weight too heavy in the overall equation:

  • Compared to the "true" amount of actual p2p solutions, the amount of "Support already helped me"-type of solutions, as well as the "quoting-other-user" answers usually is rather low. In my experience, this only occured max. 3-4 times per month (in a community with a few thousand posts/month). How often would you say this happens in your community?
  • Moderator answers (ideally) should only be posted when nobody but CS is able to provide an answer. While the number of these cases can vary due to reasons external to the community, it should stay rather consistent over time. How is your policy regarding Moderator replies / answers?
  • I am not sure if you mean FAQs when you refer to service articles. However the result would be the same - a customer would help another customer, even if he is only looking for information. E.g. if a user links to an FAQ that lists all supported SD card types, I think it is just as good as if he would list all the SD card types himself.
In the end, it will be impossible to track this metric 100% accurate. However in my opinion, the current metric is at least a somewhat consistent way to track trends in p2p support.

I personally would even go so far to say that views on a topic can be counted as p2p help, even though passive: If a user finds that they question has been asked before and reads it, to me this is also p2p help. How is your opinion on this?

Cheers,

Julian
Userlevel 7
Badge +4

Compared to the "true" amount of actual p2p solutions, the amount of "Support already helped me"-type of solutions, as well as the "quoting-other-user" answers usually is rather low. In my experience, this only occured max. 3-4 times per month (in a community with a few thousand posts/month). How often would you say this happens in your community?



We don't track this, but in about 12% of our marked answers the topic starter mentioned how he solved the problem using other means or service channels. I'm not sure if our community is average in this sense or not.



Moderator answers (ideally) should only be posted when nobody but CS is able to provide an answer. While the number of these cases can vary due to reasons external to the community, it should stay rather consistent over time. How is your policy regarding Moderator replies / answers?


Our policy is that the moderator team takes a look at all topics and decide if their help, interventions or additions are necessary. An "only answer when nobody but CS can answer" policy would probably hurt our NPS. We already often see that P2P topics are appreciated less by customers. And members often say "I want to know the company's input on this". We don't want to ignore those members.


I personally would even go so far to say that views on a topic can be counted as p2p help, even though passive: If a user finds that they question has been asked before and reads it, to me this is also p2p help. How is your opinion on this?


Page visits tell only part of the story, in my opinion. We use a survey to measure our indirect answers and calldeflection 🙂 It's probably our most important metric, seeing as 98% of our forum visits are from non-members. We also analyse the noncontact ratio: how many visits don't end up looking for the contact page of Telfort when visiting a page on our site.
Userlevel 6
Badge +3
Hi @Jurgen ,

Could you please elaborate a little bit more on how do this?
"We also analyse the noncontact ratio: how many visits don'tend up looking for the contact page of Telfort when visiting a page on our site."

This is really good 😀
Userlevel 7
Badge +4
Sure!

Non contact ratio (NCR) is a calculated metric that you can create in the well known analytics suites. I'm not sure Insided offers this out of the box, but if you enable advanced web analytics for your community you can track it :)

1. Create an audience segment of visits that contains a contact view called "Contact"
2. Create a calculated metric of '1 - ("Contact" / Visits)'
3. Now you have NCR, which you can look up for any page on your site. The higher the percentage, the better.

Segment:
https://support.google.com/analytics/answer/3123951?hl=en
https://marketing.adobe.com/resources/help/en_US/analytics/segment/

Calculated metric:
https://support.google.com/analytics/answer/6121409?hl=en
https://marketing.adobe.com/resources/help/en_US/analytics/calcmetrics/

Reply