Digital Twin Rollout Puts Measurable Factory Gains First

Unilever is scaling its use of digital twins from five factory sites to 40 locations worldwide over the next 18 months, in partnership with consulting and technology services firm Accenture. Vicky Cuthbert, chief product supply chain officer for the company's personal care business, told Supply Chain Dive that early work was probably not clear enough on making sure the company was driving a benefit that could be tangibly measured in the results. That admission is the most useful part of the disclosure, because it identifies the failure mode that stalls most industrial digital programmes before they reach scale.

Key Conclusions

First, the binding constraint on industrial digitalisation is measurement design, not technology availability. Digital twins simulate a production line so that changes can be tested before physical modification, which is genuinely valuable, but value only becomes visible if the baseline and the target metric are defined before the model is built.Unilever will carry a key lesson to each location: set clear initial goals and metrics, and track progress closely（Unilever puts business results first in factory digital twin rollout）. Programmes that start with the model and look for benefits afterwards tend to produce impressive demonstrations and unquantified returns.

Second, scaling from five sites to 40 requires repeatable process rather than bespoke engineering. Each site has different equipment, product mix and labour arrangements, so the transferable asset is the method: a standard way to define scope, model the line, validate against real output and hand over to site teams.the programme prioritises measurable gains, repeatable processes and workforce training（Unilever puts business results first in factory digital twin rollout）. Companies that treat each deployment as a unique project end up with 40 different models and no way to compare results.

Event Background And Programme Design

Why the pilot phase underdelivered

The pattern Cuthbert describes is common. Early digital twin work typically focuses on fidelity, how accurately the model represents the physical line, because that is the technically interesting problem.we were probably not clear enough on making sure that we were driving a benefit that we could tangibly measure in the results, Vicky Cuthbert told Supply Chain Dive（Unilever puts business results first in factory digital twin rollout）. Fidelity without a target metric produces a model that is accurate but not decision-relevant. The correction is to reverse the sequence: name the decision the model will inform, then define the metric that decision should move.

What the Gandhidham site contributed

The company's Gandhidham personal care manufacturing facility in India helped lay the groundwork for the wider expansion, according to the report.Unilever's Gandhidham personal care manufacturing facility in India helped lay the groundwork for the expansion（Unilever puts business results first in factory digital twin rollout）. Early sites matter disproportionately because they generate the template, the training material and the internal credibility needed to secure funding for the next tranche. Site selection should therefore prioritise transferability of the production process over the prestige of the location.

Best Practices

Define the metric before building the model

The most transferable practice is a strict sequence: identify the operational decision, name the single metric it should improve, agree the baseline with finance, and only then begin modelling. Typical candidates include changeover time, first-pass yield, energy per unit and unplanned downtime.Target's myDay employee app now has a customizable home screen and will soon include AI-backed insights（Target upgrades its store operations app for employees）, and retail operations teams are adopting the same discipline, tying tooling investments to specific store-level metrics rather than to general efficiency claims. Without a named metric, digital investment cannot be defended at renewal.

Budget for training as a first-class cost

Digital twins change how supervisors and engineers make decisions, which means training is not an optional add-on.the programme prioritises measurable gains, repeatable processes and workforce training（Unilever puts business results first in factory digital twin rollout）. Practical steps include pairing each model with a named site owner, running the model in parallel with the physical line during a validation window, and documenting the deviations so the next site starts from a better baseline. Where training is underfunded, models are built and then abandoned once the central team moves on.

Common Mistakes

The first mistake is chasing model fidelity without a decision to inform, which produces accurate simulations that nobody uses. The second is measuring against an unclear baseline, so any improvement is contestable and funding is withdrawn at the next budget cycle. The third is treating each site as a bespoke project, which prevents comparison across sites and destroys the compounding effect that makes multi-site rollouts worthwhile.

Outlook For Industrial Digitalisation

Three factors will determine whether the 40-site programme delivers: the discipline of the metric definition, the quality of the training and handover, and whether results are reviewed on a fixed cadence rather than at the end of the programme.the company will partner with Accenture to roll out the capability across its global factory network（Unilever puts business results first in factory digital twin rollout）, and external partners are useful precisely because they bring a repeatable method from other clients. The risk is that the method is imported without the internal ownership needed to sustain it.

A second consideration is how supply chain digitalisation is valued more broadly.a running tally of major retail acquisitions, mergers and IPOs tracked through 2026 provides a benchmark for deal cadence and valuation sentiment（The running list of retail deals in 2026）, and supply chain capability appears repeatedly as a rationale in transactions, which suggests buyers assign it real value. Manufacturers that can demonstrate audited savings from digital programmes are better positioned to defend investment during cost-cutting cycles.customs authorities are weighing earlier filing deadlines for low-cost imports as postal and de minimis rules continue to change（CBP eyes earlier filing deadlines for low-cost imports）, and regulatory change adds another reason to keep supply chain models current rather than static.

Summary

Unilever's expansion of digital twins from five sites to 40 shows that the limiting factor in industrial digitalisation is measurement design, not technology. The two transferable lessons are to define the metric before building the model, and to budget for training as a first-class cost.the company is scaling digital twins to 40 sites worldwide while prioritising measurable gains, repeatable processes and workforce training（Unilever puts business results first in factory digital twin rollout）. Programmes that reverse that order tend to produce impressive demonstrations and unquantified returns.