Business TopicsEnterprise SoftwareGuest BloggerOrganizationProduct ManagementProduct Marketing

Guest Post: To Kill a Product: Why, When and How, Part 1/3

Note: This is the 1st of a 3 part series of articles by guest blogger Chris Brown. If you feel inspired to write a guest post of your own, click here to find out how to submit it to us.

To Kill a Product, part 1: Why?

It’s there somewhere, in the quarterly business review, a line item that shows a couple dozen customers, down from a year ago, which was down from the year before. It brings in a little revenue, probably less than a percentage point of total, but revenue being revenue, why not keep cashing the checks?

“It,” of course, is the Product That Wouldn’t Die, its longevity no longer due to any real practical purpose but to management’s reluctance to pull the plug. But pull the plug they should. Even if costs are negligible (and sometimes they’re not), ineffective products divert focus and resources from core and growth products, and ultimately dilute the overall value proposition of the business.

The role of the product manager includes performing the kill analysis – thoughtful, thorough and completely unbiased – and, if necessary, making a recommendation that is best for the business.

Why the hard decision might be the right decision

Why should an organization kill a product, particularly if it generates even a small amount of incremental revenue? The first reason should be obvious: The product isn’t profitable. But products that are clearly bleeding money are not what we’re really discussing here. The harder decisions have to be made for products that have some revenue, some customers, have low or justifiable costs, but no long-term plan, prospects or strategic relevance.

Here are some examples of how a moribund product can have a draining, distracting effect on an organization and its customers:

On the Product team: Product managers and analysts are required to field customer feedback, analyze market data and track KPIs for a product that most likely represents a tiny fraction of the company’s business, and takes focus off of supporting the profitable products and, more importantly, new product discovery and innovation. While the beleaguered product manager is trying to keep a past-its-prime product on life support, her competition is dreaming up the “next big thing.”

On Sales and Marketing: The sales team needs to be trained on a product that very few customers actually use. Often dying products are more complicated than they’re worth (hence, why they’re in decline) and therefore can require a disproportionate amount of time and knowledge to effectively sell and support. Marketing, in addition, needs to account for these products in their collateral and messaging, adding not only to expense, but to the clutter they’re continually tasked to cut through.

On Customer Support: The best customer support teams are fluent in all the products in your company’s portfolio, even those with a small number of customers. But customers of troubled products can often fill up a Customer Support’s team queue, generating a disproportionate number of phone and email support tickets. Meanwhile, customers of more profitable products are left on hold or waiting for their email response.

On Technology: Not only can there be hard technology costs associated with supporting a product, such as server and bandwidth expenses, but also the time required to maintain both internal and external (e.g. customer-facing) processes. After all, if a product breaks, even one with relatively few customers, someone has to fix it.

On customers: With customers there’s a risk that the poorly performing product becomes a customer’s focal point, and that its performance, presumably poor, will be used as ammunition in contract negotiations or, worse, as a proxy for all the company’s products. This can present considerable sales challenges and erode the company’s reputation.

All of these distractions have associated costs, which can be harder – though not impossible – to quantify. If a kill decision is particularly contentious, it may be necessary to put a dollar amount on these costs, which can be calculated by multiplying the number of hours spent supporting the product by the estimated hourly salary of the individuals doing the supporting.

The most important reason why ineffective products need to be killed, though, is because they dilute the company’s overall value proposition. If you can measure the total value your products deliver (and you should), perennially poor performing products will naturally drag down the sum of that equation. And with that downward pressure on value come declines in employee morale, confidence in senior management, and customer loyalty. No one wants to manage a dying product. No one wants to sell, support or, certainly, buy a dying product, either. To preserve its overall value, focus resources on core initiatives and customers, and maintain a vibrant workplace, a company should be willing and able to quickly put underperforming products to pasture.

–  Chris Brown

Chris is vice president of product management at Apartments.com, a division of Classified Ventures, LLC. Email him at cbrown@apartments.com or follow him @Brown784

Coming up in this series:

Part 2: When is it time to kill a product?
Part 2 offers up six areas to keep an eye on for telltale signs. It’s examining these areas that will help product managers build the case to kill or keep a product.

Part 3: How do you kill a product?
You’ve made the decision to pull the plug, now follow these steps to ensure a smooth sun-setting process.

  1. Mike Boudreaux

    One of the problems with quantifying associated costs is that Finance will typically factor in a cost allocation when creating the product’s income statement. The most common method for doing this is to take the total costs of each of the groups (Product management, Marketing and sales, Technology, Customer support) and then allocate these costs based on the weighted revenue by product. And so, if your old decaying product has diminished sales then the allocated costs will be diminished as well – even if the product is still weighing down your organization. Maybe, by luck your allocated time will be correlated with the revenue, but probably not.

    You can create your own P&L estimate based on different assumptions, but in many companies this can create major problems with management if you don’t use consistent numbers from what they’ve seen before. Even if you get buy-in and support from Finance, be ready to back up your assumptions and bridge the gap between reported figures and your analysis.

    There can also be a great deal of emotion tied to killing off products. Make sure you know the political landmines and the history of your executive team. It could be that the product is the baby of your boss’s boss’s boss and that’s how he got where he is. Also consider other stakeholders such as a services manager who is making a killing supporting the dying product and who will want to protect their P&L metrics.

    Usually, the customers that are buying your decaying product are existing customers who can’t bear the switching costs to move to the next generation product. They might even be buying your old junk AND your new stuff. Before euthanizing the old junk, you should consider how to migrate your installed base to the new stuff.

    I’m looking forward to reading Parts 2 and 3. This is a great topic.

  2. Chris Brown

    Thanks, Robin!

    Mike, great points. I definitely agree that weighted costs allocations are not the way to go. Performing your own analysis may be politically perilous, but thorough, conservative, well-documented assumptions should hold up. All you can do is put your best foot forward and hope the executive team appreciates the effort.

    And re: the emotional ties, this is absolutely the biggest hurdle I’ve seen. In Part 3 I talk about how communications is tantamount, and how making sure all stakeholders know the full set of decision-making criteria – that this is not a random or cold-blooded move, and that there will be tangible benefits – will ensure a smoother sun-setting.


  3. Dheeraj

    Any product’s versions are a result of feature triages putting the “must-haves” in the latest version. Thus, the “nice-to-have” features are in the works while the sales and support teams are grappling with all sorts of feedback and tickets. Yet another team that is busy tweaking its services is the support infrastructure team (hardware, security, general admin etc.).

    With these variables in play, when a product / product line falters, it becomes difficult to really pin point the cause to a failing product or failing processes. It is also difficult to say if the processes are failing owing to the nature of the product.

    Why is it difficult ? Everybody is emotionally attached to what they have done and even leadership vision gets clouded.

    This has been one of the obstacles I have seen in the phasing out of products. While some members of the team are able to see that the problem originated in the envisioning of the product or that version, belling the cat is a step that sees a lot of procrastination.

    Oftentimes, sales and the product manager take a lead. These folks have to be really dispassionate.

  4. deepak

    Some time companies kill their existing blockbuster product although they are making money and mkt captalization because their positioning of product is shifted drastically. one of the example is Toyta stop their production of Qualis in india although that time Qualis is market leader and shifted from SUV segment to Passenger Car segment.

  5. Chris Brown


    It’s an interesting challenge. You want product managers to be passionate about their products, yet I agree they have to be able to maintain some distance. They need to be able to cast a clear, unbiased eye on the situation.

    That’s where the KPIs can be helpful. If the metrics you follow are well chosen, then they’ll tell the story pretty clearly.

    Thanks for the comment!

Leave a Reply

Your email address will not be published. Required fields are marked *