In 2023, I blogged about a provocative “suppose”: What if Google included an AI chatbot to Browse, effectively cannibalizing itself?

Fast-forward to 2025, and it’s no more theoretical.

With AI Overviews (AIOs) and AI Setting, Google has actually not just eaten into its very own search item, yet it has additionally taken a huge bite out of publishers , also.

Cannibalization is normally framed as a threat. However in the appropriate circumstances, it can be a growth driver-or also a survival technique.

In today’s Memo, I’m taking another look at product cannibalization with a fresh AI-era lens, consisting of:

• What cannibalization really is (and why it’s not always negative).
• Famous instances from Netflix, Apple, Amazon, Google, and Instagram.
• How Google’s AI shift fulfills the meaning of self-cannibalization– and where it doesn’t.
• The 4 big advertising and marketing ramifications if your brand cannibalizes itself in the AI-boom landscape (for premium subscribers).

Today’s Memorandum is an upgraded version of my previous guide on item cannibalization.

Formerly, I covered exactly how adding an AI chatbot to Google Browse would suggest that Google would be cannibalizing itself– which just a few companies in background have successfully completed.

In this upgraded Memorandum, I’ll walk you via successful item cannibalization examples while we revisit exactly how Google has actually cannibalized itself with a freshened lens.

Since … Google has actually successfully cannibalized itself with the consolidation of AI Overviews (AIOs) and AI Setting, yet they have not located a means to monetize them yet.

And publishers and brand names are enduring therefore.

So who wins here? (Does any person?) Only time will tell.

Product cannibalization is the substitute of a product with a new one from the exact same firm, typically shared in sales revenue.

Even though most interpretations state that cannibalization takes place when earnings is level while 2 items profession market share, there are a number of instances that show revenue can expand as a result of cannibalization.

Item cannibalization, or market cannibalization, is commonly seen as something poor– however it can be good or even required

Allow’s take into consideration a couple of instances of product cannibalization you’re most likely currently knowledgeable about:

• Hardware.
• Retail.
• SaaS/Tech.

Equipment business, for instance, need to bring out much better and newer chips on a regular basis. The lifecycle of AI training chips is often much less than a year due to the fact that new designs and higher handling capacities swiftly make the previous generation out-of-date.

Today, chips are the best product in technology– companies developing and training AI models require them in enormous amounts, and as quickly as the next generation is released, the old one loses most of its worth.

Therefore, chipmakers are compelled to cannibalize their very own products, progressing layouts to remain competitive not only with rival suppliers yet additionally with their very own previous developments.

But there are plain differences in between cannibalization in retail and technology.

Retail cannibalization is driven by seasonal modifications or customer trends, while tech item cannibalization is mostly an outcome of progress.

In fashion, for instance, customers prefer this year’s collection over last season’s. The new collections cannibalize old ones.

In technology, brand-new technology leads firms to change old items.

New PlayStation gaming consoles, for instance, substantially change sales from older ones– specifically considering that they’re backwards compatible with games.

One more instance? The growth of the brainless content administration system (CMS), which progressively changes the combined CMS and presses material monitoring service providers to use new items and features.

Netflix made numerous item pivots in its history, yet 2 stand apart one of the most:

1. The button from DVD rental to streaming, and
2. Subscription-only memberships to ad-supported income.

On November 3, 2022, Netflix introduced an ad-supported plan for $ 6 99/ month on top of its standard and costs strategies. (It has because boosted to $ 7 99/ month. See picture below.)

During the pandemic, Netflix’s customer numbers skyrocketed, however they came back to planet like Falcon 9 when Covid receded: Go into the “Basic with ads” subscription that promoted retention.

An additional obstacle for Netflix? Competitors. Lots of them– and with legacy media histories.

Initially, the technique of creating initial content and making the experience seamless throughout many nations resulted in strong growth.

Yet when competitors like HBO, Disney, and Paramount launched similar items with initial material, development reduced.

When Netflix released the ad-supported strategy, just 0. 1 % of existing individuals made the switch, yet 9 % of new customers selected it (see listed below).

A check out other streaming platforms suggests the share will certainly enhance with time. Below’s a quick look at percents of clients on ad-supported strategies throughout platforms:

• Hulu has 57 %.
• Paramount+– 44 %.
• Peacock– 76 %.
• HBO Max– 21 %.

However, Netflix’s brand-new strategy is not technically thought about product cannibalization yet partial cannibalization based upon price.

The product is the same, yet with the new plan, it’s now available to a brand-new client sector that formerly wouldn’t have considered Netflix.

In addition, it prevents an existing client section from churning considering that the economic crisis shuffles the investing actions of client segments

We can wrap up that the brand-new ad-supported Netflix strategy is not the same sort of cannibalization as its streaming service

In 2007, internet links became solid sufficient to open the door to streaming. Netflix was not the very first company to provide movie streaming, yet the initial one to be successful at it.

The company paved the way by incentivizing DVD service customers to involve online, for example, with rental lines on Netflix’s site. However inevitably, the pivot was the outcome of technological progression.

Another item that saw the light of day for the first time in 2007

The apple iphone.

When it launched, the iPhone had all the functions of the iPod and even more, making it a situation of full item cannibalization

Because of this, the share of income from the iPod substantially decreased once the apple iphone released (see image below).

Although you might say it’s a “normal instance” of market cannibalization when looking at revenue streams from each product, it was a technical step-change instead of partial cannibalization based upon prices.

Nevertheless, huge action in technology don’t always cause a preferred substitute of an old item.

Take the Amazon Kindle, for instance.

Released in 2007– just like Netflix’s streaming item and the iPhone (something was up that year)– Amazon brought its new ebook reader, Kindle, to market.

It made such an impact that individuals anticipated the death of paper publications. (And curators everywhere giggled while booksellers supported themselves.)

However over 10 years later on, ebooks supported at 20 % market share, while print publications recorded 80 %.

The factor is that publishers got involved in pricing fights with Amazon and Apple, which additionally started to play an important function in the e-book market. (It’s a lengthy story; however you can check out it right here

Amazon tried to cannibalize its core service (books) with the Kindle (books), but couldn’t make item pricing work, which caused e-books frequently being extra costly than print versions. Yikes.

The innovation transformed, yet consumers weren’t incentivized to use it.

Let’s look at 2 final examples below. These 2 firms gotten or duplicated competitors to control partial cannibalization:

• YouTube video clips are technically far better answers to numerous search inquiries than internet results. Google saw this really early and smartly acquired YouTube. Video results spent some time to fill even more area in the Google SERP, even though they practically cannibalize web outcomes. Yet today, they’re typically several of one of the most aesthetically impactful results (and commonly the first outcomes) that searchers see.
• Instagram saw the success of Snapchat stories and chose to duplicate the function in order to reduce competitor development. Regardless of the cannibalization of normal Instagram blog posts, internet interaction with Stories was greater. (And talking YouTube, you might suggest that YouTube Shorts comply with the very same concept.)

With all this in mind, we can state there is complete and partial cannibalization based upon how many attributes a new item replaces

Rates adjustments, duplicated features, or purchases result in partial cannibalization that doesn’t cause the exact same earnings growth as full cannibalization.

Full cannibalization calls for 2 problems to be true:

1. The brand-new item has to be built on a technical action adjustment, and
2. Customers need to be incentivized to use it.

With this expertise foundation in place, allow’s check out the changes in the Google Search item over the last 12 – 24 months.

Let’s use these product cannibalization concepts to the instance of Google vs. ChatGPT & Co.

In the original web content of this memo (released in 2023, I shared the following:

If Google were to add AI to Browse in a similar method as Neeva & Co (see previous post regarding Early tries at integrating AI in Search , the list below conditions would certainly hold true:

1. AI Chatbots are a technical step-change.
2. Customers are incentivized to utilize AI Chatbots since they offer fast and good solution to most concerns.

Nevertheless, not all conditions are true:

1. AI Chatbots don’t supply the complete capability of Google Search.
2. It’s not cheaper to incorporate an AI Chatbot with Browse.

I’ve been clear about my hypothesis for some time now. As I shared in my 2025 Halftime Record :

I directly think that AI Setting won’t introduce [fully in the SERP] prior to Google has found out the money making design. And I anticipate that searchers will certainly see way less advertisements but better ones and showed at a much better time.

And I highlighted this in Is AI cutting into your conversions? also:

Google won’t show AI Setting all over, due to the fact that adoption is generational (see the UX research study of AIOs for more details). I believe AI Mode will certainly go for a more comprehensive scale (like turning up for even more inquiries generally) when Google determines monetization.

And also, ChatGPT is not yet monetizing, so advertisers go to Google and Meta– in the meantime. And that’s my theory regarding why Google Look is continuing to expand.

Remember, to effectively cannibalize your existing item, you need customers to wish to utilize it. And according to a recent report from Garrett Sussman over at iPullrank, over 50 % of individuals who attempted Google’s AI Mode when and didn’t return. [Source] (So it appears Google’s still identifying the incentivising component.)

Despite the advancements we have actually seen in the last 6 to 12 months with AI designs– and the incorporation of live web search and item suggestions into AI talks– I would certainly suggest that they serve for information-driven or generative inquiries but lack the data sources needed to offer great answers to product or service searches.

Allow’s have a look at an example:

If you input “best plumbing in Chicago” or “ideal toaster” right into ChatGPT, I would certainly say you ‘d actually obtain much less top quality results– in the meantime– than if you input the exact same queries into Google. (Go try it for yourself and allow me understand what you locate. Yet here’s a walk-through with Amanda Johnson hopping in to illustrate this listed below.)

At the same time, these product or services inquiries are the inquiries that online search engine with an ad revenue company model can generate income from ideal.

It was stated that ChatGPT expenses at least $ 100, 000 each day to run when it initially went across 1 million customers in December 2022 By 2023, it was costing concerning $ 700, 000 a day. [Source]

Today, it’s most likely to be a significant multiple of that.

Bear in mind, Google Browse sees billions of search queries every day.

Despite having Google’s advanced framework and skill, AI chatbots are expensive

And they can (still) be slow-moving– despite having the developments they have actually made in the last 12 months. Current and traditional Google Look systems (like Included Fragments and Individuals Also Ask) may offer a much faster response.

Yet, alas, below we are in 2025, and Google is cannibalizing its very own item via AIOs and AI Mode.

Today, according to Similarweb data, use of the AI Mode tab on Google.com in the U.S. has somewhat dipped and now sits at simply over 1 %. [Source, Source]

Google AIOs are now seen by more than 1 5 billion searchers every month, and they sit front and facility. However interaction is dropping. Individuals are investing less time on Google and clicking fewer pages. [Source]

But Google has to compete with not just various other online search engine that supply an AI-chat-forward experience, however also with ChatGPT & Co. themselves.

Listed below, I’ve detailed out essential considerations for your brand name if you might use item cannibalization as an approach.

You’ll want to:

1. Reframe cannibalization as a calculated alternative for the brand as opposed to a failure.
2. Make use of the complete vs. partial cannibalization lens.
3. Examine the two success conditions.
4. Protect your core offerings while you experiment.
5. Use competitive cannibalization defensively.
6. Screen, discover, and adjust.

In the next area, for costs customers, I’ll walk you via what to watch out for if you determine to use product cannibalization as a growth approach.

1 Reframe Cannibalization As A Strategic Option

• Don’t fail to seeing item cannibalization as a failure; analyze if it can safeguard market share or increase development.
• Audit your product line and GTM strategy to determine areas where you can self-disrupt prior to a competitor does.

2 Usage The Complete Vs. Partial Cannibalization Lens

• Full cannibalization works best when there’s a tech leap and strong client incentives.
  • Instance: Apple iPhone changing iPod– all iPod features plus far more capability brought about the iPod’s rapid decrease.
• Partial cannibalization through rates, functions, or purchases is much less high-risk yet might not supply big growth.
  • Example: Netflix ad-supported strategy– same streaming item, however a lower-cost rate opened the door to brand-new sections and lowered churn risk.
• Map present and future offerings against these two classifications to determine your strategy.

3 Test The 2 Success Problems

A cannibalizing item is more likely to prosper when both are true:

• Tech Jump: Provides a meaningfully much better means to resolve the issue.
  • Example: Netflix DVD → Streaming in 2007 leveraged quicker internet rates to alter the distribution version entirely.
• Client Reward: Lower price, much better efficiency, even more benefit, or standing.
  • Example: YouTube procurement by Google made richer, extra visual answers possible in Browse, boosting the user experience.

If both apply → go after full cannibalization.

If one applies → go after partial cannibalization with controlled range.

4 Shield Your Core While You Experiment

• Identify high-revenue segments and shield them from very early disturbance.
  • Example: Google maintaining AI Setting away from highly monetizable queries like “finest bank card” till the advertisement version is ready.
• Examination self-disruption in lower-stakes markets to verify demand prior to scaling
  • Instance: Instagram Stories presented in such a way that increased web involvement while protecting the feed’s ad stock.

5 Usage Competitive Cannibalization Defensively

• When a rival launches a hazard, select between:
  • Get: Google obtaining YouTube to manage the increase of video clip as a search solution style.
  • Replicate: Instagram adopting Stories from Snapchat to quit customer movement and expand engagement.
  • Set apart: Amazon Kindle– a technology leap that attempted to move readers from print to digital, but without a cost benefit, fostering plateaued.

6 Display, Learn, And Adjust

• Track involvement, income mix, and fostering by section.
  • Example: Similarweb information on AI Mode– U.S. use holding simply over 1 %, signaling restrictions to adoption speed.
• Adjust rollout rate based on generational fostering patterns and competitor steps.
  • Example: Google AIO engagement decrease– revealing that positioning alone does not guarantee sustained customer passion.

A fine example is exactly how to do this is Chegg.

The business has been taken out by Google’s AI Overviews and also filed a claim against Google. Chegg’s value was answers to research concerns, but because nearly every trainee utilizes ChatGPT for that, their worth chain damaged. Exactly how is the company reacting to this life-ending danger?

In my Q 2 Marketplace Deep Dive , I discuss that Chegg has located a lifeboat:

Chegg has released a new device, Service Precursor, that allows trainees to compare solutions from ChatGPT & Co. with Chegg’s archive.

Rather than attempting to beat AI Chatbots, Chegg hits them where it harms: in the hallucinations.

LLMs can make things up, which is specifically uncomfortable when it concerns learning and taking examinations. Envision you invest hours internalizing the incorrect facts!

Service Precursor confirms AI solutions with Chegg’s archive of human-sourced product. It contrasts the answer from fundamental versions and highlights differences and consensus.

---

Included Photo: Paulo Bobita/Search Engine Journal