Pricing Optimization Solutions

No current Forrester Wave covers this category. The Pricing Optimization Solution Landscape, Q3 2023 maps the market without scoring it, and a scored evaluation is expected to follow. What follows draws on that research and on Forrester's earlier assessment of the market.

Forrester applied its Wave methodology to retail price optimisation around two decades ago, evaluating five vendors, and published the result under a headline stating that the solutions were still emerging.

The vendors were Khimetrics, DemandTec, Retek, i2 Technologies, and Manugistics. Three further vendors, ProfitLogic, JDA, and SAP, declined to participate. None of the five evaluated survive as independent companies today.

Twenty years later there is a Landscape and no Wave. A market that was emerging in the mid-2000s has still not settled into a form Forrester considers ready for a scored comparison.

That persistence is worth explaining, because the reason has little to do with the technology.

Why pricing resists automation

Every category in this series involves software that recommends and an organisation that does or does not act. Pricing is the case where the resistance is strongest, and the reasons are structural.

Price is the most consequential single variable a business controls. A percentage point of margin moves more profit than most operational improvements, in both directions, and the effect is immediate. That makes it the decision executives are least willing to delegate to a model whose reasoning they cannot follow.

Price is also political. In most organisations pricing authority is distributed across product management, finance, sales leadership, and regional teams, each with a legitimate claim and a different objective. Product wants positioning, finance wants margin, sales wants to close, and regional wants local competitiveness. A pricing optimisation system produces one recommendation, which means one of those parties loses an argument they were previously able to have.

And price carries relationship history. A customer who has held a rate for four years reacts to a change in a way that a model optimising for margin does not anticipate. Sales teams know this, which is why they override recommendations, and why the override rate is the metric that actually determines whether a deployment worked.

That combination is why a market with obvious value has taken twenty years to mature.

What the Landscape describes

Forrester's Q3 2023 research positions these solutions as helping maximise revenue and drive growth by accurately and optimally pricing products, and identifies several capabilities as characteristic.

Real-time price adjustment for industries susceptible to price and demand fluctuations, with Forrester naming retail, e-commerce, supply chain, mining, agriculture, and fishing among them.

Scenario simulation, letting product and pricing managers model the impact of a change before it reaches the market. Forrester describes these simulations as providing real-life scenarios of the effect of pricing changes on products sold.

And discounts, trade promotions, campaigns, and loyalty programmes, which Forrester found present as standard features across many solutions.

The simulation capability is the one worth weighting most heavily, and not for the reason vendors emphasise.

Its stated value is better decisions. Its practical value is that it converts a pricing argument from an assertion contest into an evidence contest. When product management, finance, and sales disagree about a price change, a simulation gives each party the same projected outcome to argue against, which is a considerably better starting position than three sets of intuitions.

That is a governance benefit rather than an analytical one, and it addresses the organisational obstacle described above more directly than any accuracy improvement.

Two markets under one label

Forrester notes that these solutions address both business-to-business and business-to-consumer markets, and that the two may differ in many respects.

That is an understatement worth expanding, because the difference determines which vendors can serve you.

Consumer pricing is list pricing. A price is set, published, and applied to anonymous buyers. Optimisation means finding the price point that maximises revenue or margin across a population, informed by elasticity, competitor prices, inventory position, and seasonality. The decision is made once and applies broadly, and feedback arrives quickly through sales volume.

Business pricing is negotiated pricing. Each customer has a contract, a discount structure, a volume commitment, and a history. Optimisation means guiding a specific decision about a specific deal, within an approval hierarchy, against a floor that varies by customer segment. The output is a recommended price for one negotiation, delivered to a salesperson who may disregard it.

Those are different products. A consumer pricing engine tuned for elasticity modelling across a catalogue is not solving the B2B problem, and a deal-guidance tool built around approval workflow is not solving the retail one.

The industries Forrester lists lean heavily toward the first: retail, e-commerce, and commodity-adjacent sectors where demand fluctuates and prices are published. Buyers in negotiated B2B environments should establish early which side a vendor was built for, because the category label conceals it.

The relationship with quoting

For anyone working in revenue operations, pricing optimisation sits directly upstream of configure, price, quote, and the distinction between them matters.

CPQ enforces pricing rules. It knows the list price, the discount authority, the approval thresholds, and the terms, and it applies them consistently so that a quote is valid and a seller cannot exceed their authority.

Pricing optimisation decides what those rules should be. What the list price ought to be, where the discount floors should sit, which segments justify which treatment, and what a given deal is likely to close at.

Organisations frequently conflate them, and the symptom is a well-implemented CPQ enforcing a discount structure nobody has revisited in three years. The system operates correctly against rules that stopped reflecting the market.

The integration between them is where the value compounds. An optimisation engine that recommends a price which CPQ then enforces, with the outcome fed back to improve the next recommendation, is a closed loop. One that produces a report which somebody manually translates into a revised price book once a year is not.

Where algorithmic pricing gets uncomfortable

This category carries risks that most enterprise software does not, and they deserve stating plainly.

Dynamic pricing has a reputational dimension. Consumers accept that airline and hotel prices move, and react considerably less well to price movement in categories where they expect stability, particularly essentials and particularly during disruption. Several public backlashes have followed pricing systems behaving exactly as designed during events where the design was not the problem, the optics were.

Personalised pricing raises fairness questions that are not settled. Charging different individuals different prices for the same good, based on inferred willingness to pay, is legal in many contexts and sits poorly with most people's sense of fair dealing. Where the inference correlates with protected characteristics, whether or not those characteristics were used as inputs, the exposure is more than reputational.

And algorithmic coordination has drawn competition authority attention in several jurisdictions. Where multiple competitors use pricing tools drawing on shared data or common algorithms, prices can converge without any agreement between the firms, and regulators have been examining whether that outcome falls within existing prohibitions. The area is developing and the analysis differs by jurisdiction.

None of these argue against the category. They argue for treating pricing automation as a decision requiring governance rather than as an efficiency purchase, and for knowing what your system is permitted to do without human approval.

That last question is the same one running through every agentic category in this series, and it lands with unusual force here because the action is visible to customers.

What a scored evaluation would need to settle

If Forrester publishes a Wave, three choices would define the category.

Whether B2B and B2C are evaluated together. Combining them produces a criteria set that fits neither well, and separating them acknowledges what practitioners already know.

How autonomy is treated. There is a wide gap between a system that recommends prices for human approval, one that adjusts within bounded rules, and one that sets prices continuously without intervention. Those are different risk profiles and different products, and a criteria set that does not distinguish them describes the market imprecisely.

And whether governance is scored. Approval workflow, audit trail, override tracking, and the ability to explain why a price was set are the capabilities that determine whether an organisation can defend its pricing to a regulator, a large customer, or its own board. They are unglamorous and they are the ones that decide deployments.

What a buyer can establish now

Without a scored evaluation, the useful work is internal and it is mostly about honesty.

Establish where pricing authority actually sits, including the informal version. If a regional sales director can approve an exception by phone, the system will route around itself and the recommendations will describe a process that is not the one running.

Measure your current override rate if you have any pricing guidance today. That number predicts adoption better than any capability assessment, and if it is high the constraint is trust rather than accuracy.

And decide what the system is optimising before evaluating anything. Margin, revenue, volume, market share, and customer lifetime value produce different prices, and an organisation that has not chosen between them will get a recommendation that optimises whatever the vendor's default assumed.

That choice is a strategy decision rather than a configuration setting, and it is the one that determines whether the output is useful or merely confident.

Forrester does not endorse any vendor named here, and nothing in this article should be read as a recommendation to buy. Competition and consumer protection positions described here vary by jurisdiction and change; nothing here is legal advice.