This post is for anyone who's thought:
"How likely am I to be disrupted by chat search & how bad will it be?"
It's meant to provide an approach you can take to understand the potential impact of chat search on the traffic your website currently gets from organic rankings by figuring out where you're most disruptable.
Table of Contents:
1. Understand Benefits of Chat Search Engines vs "Googling It"
By now I probably don’t need to define what “Chat Search Engines” is, but I would like to in the context of how it differs from Search as we know it today - “Googling things.”
Chatbots stay in context
Chat search engines are interactive and able to stay in context like a conversation. Go run this example yourself:
- Search: "What are the top 10 SUVs under 65,000"
- Refine: “That also has a third row” (no need to repeat the text above like you would in Google)
- Visualize: "Turn it into a table ordered by price" (here's what happens when you "Google it")
You don’t have to repeat your context over and over again, the benefit to the user is ease of getting to exactly what they need.
Searches are becoming more descriptive
With a chat based interface for your searches and refinements, I believe queries will be longer like the example above. This would be a benefit if that data is ever shared with marketers, as each word in a search phrase tells me something deeper about the user's intent.
Chat search engines display the answer with no clicks
Whatever is going to make it easier and faster for the searcher to get to their answer is going to win. I really hope we retain links to the sources I think Perplexity does this best, followed by Bing.
Chat search engines to experiment on:
2. Combine your PPC and SEO Data
Ok, now that we’ve got the basics of chat search engine value proposition out of the way, lets dig in on combining your PPC and SEO data to determine the risk that might be ahead of you for your organic and paid search campaigns, because this chat search thing just might lower the ROI of your SEO investments.
How to combine your PPC and SEO Data:
- Get all the words you pay for from your search Query report in Google Ads
- Login to Google Ads
- Create a new report
- Add: Search Term, Cost (filter out $0 cost), Clicks, Conversion
- Run them through a rank tracking tool (Nozzle, Traject, Stat are all great companies) and get that export for your organic data.
- Now you can harvest Google’s intelligence on that keyword (Like answerboxes) and join it to your paid search data to look for trends.
Disruptables: Answer Boxes & Knowledge Graphs
Google has been answering people's questions with single answers for ages, ever seen an answer box like this?
Answer boxes are a hint that Google believes this query might be answerable without you having to visit a website. I think that hint might mean that ChatGPT and other chat search engines might also provide instant answers.
Our hypothesis is that search queries that show a knowledge graph will be first up on the Chat Search disruption list.
Analysis: Trend how often you're showing the knowledge graph on high-value keywords
Below shows a month-over-month trend of percent of paid search terms that are:
(1) Showing in the Top 3 ranking AND (2) Trigger an answer box SERP feature
This is a screenshot out of our cloud-based data platform, Supernova™.
For client #6, we'd likely advise:
- We looked at all keywords where an answer box is in the top 3.
- Google was showing answer boxes on 25% of your total conversions this month 53% of your conversions were on queries that showed an answer box. If searching via chatbot is .1% of your searches no big deal, but imagine if it is 10% or 25% of searches someday.
- Using Supernova (our tool to combine data across typically siloed data) we are able to set alert thresholds for each client, so if the client is thinking, tell me when I am at 50% then we would have triggered the email and alert to the team member, their manager, etc to have a conversation about what to do, now that we crossed a threshold the client cares about. No need alerting a client to a metric that isn't big enough to take action or shift strategy.
Let's take this search for Paralegal - “Googling it” gives us an answer box.
Google has learned not only that a simple answer is a good one, but they also have a People Also Ask feature below that, indicating they understand what the searcher is likely to ask next.
Now is you execute that search with a chatbot like ChatGPT, you can see that I got a robust answer and it also proactively answers some of the questions that Google has in People Also Ask (specifically the main role of a paralegal) without me having to click to expand all of them, and then have to click on a bunch of different sites:
Now you see why I believe these kinds of searches are most likely to be disrupted.
Going back to the basics we established on the benefits of chat search in the beginning of this post - whatever is going to make it easier and faster for the searcher to get to their answer is going to win. And right now, Chat GPT-3 is giving a no-click answer that also anticipates your next question.
These are your canary in the coal mine warnings, so to speak.
Chatbot assisted search, going to tank traffic?
For my advanced data-loving SEO’s - you could also layer in your CTRs (Click Through Rates) and Impressions on your ranking URLs when answer boxes shows up to also get a feel with your own data on how answer boxes showing up impacts traffic.
Then share with your executives and leaders your prediction on how chat search - as it reaches a 5%, 10%, 30% threshold - could impact traffic. You’d likely have to join Google Search Console (GSC) data to Rank Data at the URL level and then join the Rank Data to your traffic data at the URL level.
Yes, the ROI of content might decline if chatbot driven search takes off.
Got analytics, track the referrals and browsers:
- Track the referral string for Bing's chatbot search.
- Keep an eye on Edge browser activity by month (since Bing is making you use Edge to access their chatbot).
- Bing said they will share some info in Bing Webmaster Tools next month.
Some PPC Implications of search via chatbot
Do answer boxes impact CTR?
I’m very early in my thinking on this, but in a world where you go from 20% of your cost in paid search showing an answer box to 45% in 2 months, I can’t help but wanna know if more answer boxes showing up impacts my CTRs in paid and ability to convert?
So I would like to see these trends side by side.
Where will the ads go in chatbot search?
Then there’s the battle for search ad real estate and Microsoft is willing to pretty much go negative on gross margins to take market share, and their chat search powered by OpenAI / ChatGPT seems to be quite a good bit better than Google's Bard.
This is their opportunity they’ve been waiting for - the playing field has been leveled and their CEO is ready.
I don’t know how Microsoft is going to deploy ads in chat search - this early attempt definitely looks a bit heavy handed in a chat search full of text getting images like this was a bit jarring, but they’ll figure that out.
Will chat search drive less real estate for ads? I'm guessing yes.
And ads are a large part of the Google playbook to grow revenue - add more ads, get rid of pages so you keep scrolling (see more ads), get rid of the colors around the ads, etc.
We’ll see, but for now, this searching via chatbot thing is just what our industry needed, it levels the playing field between old heads like me and the folks coming up who want to use tools like ChatGPT, AutoGPT and all the tools out there.
Lets get ready yall, its gonna be a ride.
Sign up for our newsletter to keep up with our learnings as this industry shake-up evolves: