A backtest is a historical simulation generated from a set of assumptions. It can be extremely useful to evaluate an investment idea, compare factor definitions, explore a thematic universe or test weighting constraints. An index, however, is something more demanding. It is a formally specified methodology designed to be calculated consistently through time, through future rebalances, corporate actions, data issues and changing market conditions.

That distinction is central to the self-service indexing era. The objective should not simply be to produce more performance curves, but to create methodologies that can be operated, monitored and maintained over time.

A performance curve is not an operating methodology

At first glance, many systematic strategies appear straightforward to reproduce: select a universe, apply filters, rank securities, assign weights, rebalance periodically and calculate historical returns. This may be enough to generate an attractive chart, but it is not enough to operate an index-grade strategy.

As soon as a strategy becomes a benchmark, an investable index, a model portfolio or a continuously monitored internal benchmark, additional questions emerge. Which data were actually available at each review date? How are missing observations, multiple listings or liquidity constraints handled? What happens when a company pays a special dividend, completes a spin-off, is acquired or delisted? How are cash balances, FX conversion, withholding taxes and dividend reinvestment treated? Is the methodology calculated as Price Return, Gross Total Return or Net Total Return? How are methodology changes documented and versioned?

A historical simulation can leave many of these questions implicit. An index cannot. To be reproducible and operational, every important assumption must be explicitly defined.

An index must survive the market

The real test of an index methodology is not whether it would have performed well in the past, but whether it remains unambiguous when confronted with real-world events. An index must process scheduled reviews and rebalances, react consistently to corporate actions, handle data corrections and exceptional situations, and continue to be calculated and monitored through time.

This is why an index cannot be reduced to a signal, a ranking model or a weighting formula. Selection rules and weighting schemes are important, but so are calculation conventions, corporate-action policies, execution assumptions, data governance, controls, logs and auditability.

In practice, the object being managed is not simply a portfolio idea. It is a methodology combined with an operating model. Without that operational layer, a strategy may be interesting from a research perspective but difficult to maintain as a long-term benchmark.

Why this matters in self-service indexing

If self-service indexing only makes historical simulations easier to generate, it addresses part of the research challenge but not the institutional one. It may even create a new source of fragmentation: hundreds of customized strategies whose historical results appear credible, yet whose methodologies are insufficiently specified to be reviewed, reproduced or operated consistently over time.

The promise of self-service indexing should therefore be more ambitious. It should enable institutions to create methodologies that are explicit, transparent, documented and capable of moving seamlessly from research to recurring calculation. Customization at scale requires methodological discipline at scale.

This is also why the bridge between research and production is so important. In traditional workflows, a strategy may be prototyped in a notebook, tested in a research environment, manually described in a rulebook, reimplemented in a production system and monitored through separate operational processes. Each handoff introduces translation risk. A modern index-technology stack should instead represent methodologies as structured, inspectable and versioned objects that can power simulation, analysis, documentation and live calculation from the same source of truth.

What this means in Folio

This principle is central to the architecture we have built at MCFT. In Folio, methodologies are captured through explicit structured configurations rather than hidden scripts. The platform's portfolio-based calculation framework models an index as a rule-driven portfolio evolving through time, with positions, prices, currencies, cash balances, corporate actions and execution events handled within a single coherent framework.

Folio supports both rapid simulation and key index-grade requirements, including the treatment of dividends, splits, M&A events and delistings; Price/Gross Total/Net Total Return as well as currency-hedged, decrement, volatility target or other overlay calculations; configurable rebalancing assumptions; automatically generated methodology documentation; and recurring live calculation supported by monitoring tools and operational controls.

Most importantly, simulation and live calculation rely on the same methodological foundations. A strategy can move from idea to simulation, from simulation to review, from review to documentation and from documentation to recurring calculation without being rebuilt at each stage. This continuity is what allows self-service indexing to evolve beyond backtesting and become a genuine operational framework.

The real question

The key question for self-service indexing platforms is therefore not: How quickly can a user generate a backtest?

It is rather: How reliably can a user create a methodology that remains transparent, reproducible and operational through time?

The MCFT Founders - Romain Charlassier, Jonathan Klein and Lucas Mouilleron