Clicks and open rates as metrics: how accurate are they?

[Charlotte_Lfg]

Hi there, 

We'd like to use clicks and opens as a metric for scoring, but I'm not sure how reliable those metrics are. We activated the bot activity identification, but it looks like there are still some 'fake' clicks (i.e. people click before the mail is open) which I assume is the security testing the mail. I'm not sure that the open rates are by real people. 

How do you do to use those metrics? Or do you completely disregard them? 

[Darshil_Shah1 · Accepted solution]

This is a great question. So, as you're noticing, even with the bot filtered turned on, there's still always a possibility of bot activities creeping in. Marketo keeps monitoring for such false positives and updates its bot filtering algorithm periodically so it is accurately filtering bot activities. Having said that, over the years, I've seen users giving less priority to opens/clicks. I see some users still include those as a part of their scoring model, but, I haven't seen users fast-tracking a person to MQL or inflating their scores by leaps and bounds just based on a click/open, which is fairly reasonable IMO.

[Charlotte_Lfg]

Thanks Darshil for the reply!

What would you recommend for our scoring model then? We have a very limited amount of channels, so if we exclude those, there's not much left for scoring... 

[Darshil_Shah1]

Apart from clicks and opens below are a few interactions that people use in their scoring model to give you an idea. These are just high-level items- an actual scoring model consists of many sub-divisions of the following based on the different types of events, offers, etc.

1. Events - Register, Attend, Attend-On-Demand, etc.
2. Access content offers Download (gated/non-gated)
3. Fills out form
4. Trial/Demo Requests
5. Web visits on high-value web pages (for example- pricing, renewal, etc.)
6. Viewed multiple web pages in a given time (say a day or two)

Additionally, I'd also recommend thinking around the lines of deducting points for in-activity or activities that signal a person may not be an ideal prospect, e.g., unsubscribe, visit the careers webpage, haven't reached a success step for any program for quite a while, etc.

[Charlotte_Lfg]

Hi Darshil, 

Regarding the deduction of points, this is what we already have in place - i.e. unsubscribe, status "demotion" - but as we don't have a website, nor content-gated assets, our options of interactions for scoring are rather limited. 

I'll see what I do then, thanks for answering. 

[Ruchi_Lapran1]

Hi Charlotte,

If you are attributing the channels and lead sources, they can also be included in scoring the outcomes. Even the non-gated content might have some links or CTAs which can be powered by including UTM tracking which would fetch all the relevant scoring parameters for you.

Additionally, you can also use program statuses in your scoring parameters.

[Darshil_Shah1]

---

but as we don't have a website, nor content-gated assets, our options of interactions for scoring are rather limited. 

I'll see what I do then, thanks for answering. 

---

Then, I suppose your primary interaction sources would be Email, Marketo forms and LP interactions, content syndication (your content posted on other sites, maybe), list imports, and other offline channels like tradeshows, etc. You can build your scoring around these. Just make sure you aren't double-scoring people for a single interaction- e.g., scoring on a page visit and scoring on a program status change that was triggered by a web page visit. This could easily inflate your score. I'd recommend you build a fairly consistent scoring and processes to ensure it's agile, accurate, and easy to maintain at the same time. 🙂