Help Needed: Creating KPIs for Invalid Visit Rate & Bot-Filtered Conversion Rates in Adobe Analytics

[Madhusudan_S]

Hi Team,

Has anyone implemented KPIs such as "Invalid Visit Rate", "Conversion Rate without Bots", and "Conversion Rate with Bots" in Adobe Analytics dashboards?

Additionally, I’m looking to generate a report that answers:
"How much bad bot traffic is attacking our site?"

I've attached a sample dashboard for reference to provide more context. Any guidance, best practices, or examples would be greatly appreciated.

Looking forward to your quick responses!

Thanks in advance,
Madhusudan Sura |+91-9423067216

[Jennifer_Dungan]

There is no definitive way to identify all bots... there are a lot of different ways that people try to do this, sometimes throwing out good traffic with the bad...

The first thing I need to ask you is... what is your final plan?

• Is this trying to identify bots after the fact (that got by your rules)?
• Are you going to try to identify your bots and try to have them identified by your Bot rules?

I assume from the way you are talking, it's the first... and we all know that User Agents and IP addresses are becoming less viable (we really need additional rules for Bot filtering).

So some of the things that people use to identify bots are:

• Low Resolution Usage (like 800x600 - however, you need to know your audience, we actually do have a lot of older audience using older machines, so 800x600 isn't a great indicator for us)
• Single Page Visits (this might work better for retail, however there is still going to be users that come to the site just to compare the price of something they plan to purchase elsewhere, and on publishing sites like news / blogs, one and done visits are frequent)
• Old Browsers, or strangely identified Browser Versions (similar to resolution, older audiences don't always upgrade to the latest and greatest)
• Old OSs, or strangely identifies OS versions (like above)
• Going through pages too fast (this would be a threshold you would need to determine, but sometimes you can identify bots by looking at the average time on a page... you might be able to code something in your tag manager to flag a user if they hit a certain threshold)
• The Same IP identifying as a new user / visit on each page in a session (Some bots won't allow cookies, so they trigger a new visit on each page)
• Desktop devices with no mouse (this could indicate the use of screen reader, or it could be a bot that doesn't rely on a mouse to navigate, you might want to pair this with no use of Tab and Enter keys, which are often used for navigation with assistive technology... but they could also be using voice commands)

Some of these can probably be done with segments, others would require coding to detect patterns and to flag those users, and that flag can be built into your "Bot Segment".

Then it's really a matter of stacking that exclusion segment to get conversions without Bots

[Vinay_Chauhan]

Hi @Madhusudan_S 

Here’s how I would approach delivering these KPIs in practice -

1. Turn on Adobe’s Bot Rules in the report suite (IAB standard + custom) so known bot traffic is filtered out of baseline metrics, and use the “Bots & Bot Pages” reports for visibility.
   
   

2. Build a Suspected Bot segment using heuristics (e.g. visits with no KPI events, ultra low interaction, suspicious resolutions / IP / browser combos).
   
   

3. Define your KPIs:

   • Invalid Visit Rate = flagged bot visits ÷ total visits

   • Conversion Rate w/ Bots = conversions ÷ total visits

   • Conversion Rate w/o Bots = conversions in non-bot visits ÷ non-bot visits

   • Bot conversion “leakage” = conversions in bot traffic ÷ total conversions
     
     

4. Run dashboards where you show the traffic breakdown (all vs non-bot), conversions, conversion rates, and how much bot traffic may be inflating or distorting metrics.
   
   

5. Monitor over time, and tune your bot heuristics based on real data (e.g. how many “suspected” visits overlap with Adobe’s bot filtering vs how many slip through).