A scenario
Nothing you did changes. Everything your customers are choosing between does. This is the scenario where the company is a spectator for two months and then finds out which of its accounts were staying out of habit.
Run this on your own numbersAll scenarios
Ships as a starter scenario with two levers: a competitor launch at 0.35 in month two, and a competitor price move of minus 8 percent in month three.
Mechanics
Two levers, three months apart, and then nine months of consequences that arrive through your customers rather than through your accounts.
The launch lever multiplies each competitor's quality index by one plus the launch score times that competitor's strength. At a score of 0.35, a rival at strength 0.75 improves 26 percent and one at 0.45 improves 16 percent. The market quality your customers compare you against is the strength weighted average of all of them, so a launch by the strongest rival moves the comparison most.
Your quality index does not change. The ratio does. In the churn deviation the term is minus 0.8 times the log of your quality over theirs, which means churn rises even at an unchanged price. In a named account's renewal decision the same ratio appears at weight 0.9, against price at 1.1 and switching cost at 0.8.
The second lever cuts every competitor's price index by 8 percent. Your price did not move, so the ratio of your price to theirs rises by about 8.7 percent. That reaches churn through the pressure term and reaches new business through the new customer price ratio, which carries 60 percent of the elasticity at 1.25 times.
Segments lose accounts every month. The price led segment, at 1.7 times the elasticity and half the switching cost, goes first. Contracted customers are held by the lock term and leave later, which is why the damage in this scenario usually peaks in the second half rather than the first.
Each one has a clock drawn between 40 and 130 percent of its term. A worse relative quality moves the logistic directly. Accounts that renew with a leave probability above 30 percent stay but take 3 percent off, so even the wins cost something.
If churn sits more than half again above baseline for two months, the chief executive starts spending about 1.2 percent of revenue a month on retention, which buys satisfaction, which buys retention back with a lag of a month or two. It is on a four month cooldown. Nothing else in the model answers a rival product unless you add a lever that does.
Satisfaction moves 18 percent of the way to its target each month and feeds new business through a referral term. So a quarter of elevated churn keeps costing you new logos after the churn itself settles down, which is the part that makes this scenario worse than it first looks.
The ecosystem
Every twin gets all nine classes whether or not your files mention them. In this scenario four of them do the work.
Objectives: take share when you give them an opening at weight 0.45, protect their own margin at 0.35, and not start a price war they lose. Here the launch is imposed rather than chosen, which is the point of running it as a scenario. If no competitor file was uploaded, the rivals are inferred and drawn faded in the graph.
Objective: pay as little as the job allows, at 1.7 times the base elasticity, with half the switching cost and a contract multiplier of 0.4. They feel both levers at full strength and they feel them first.
Objective: keep getting what they signed for at weight 0.4, renew at a price they can defend at 0.35. The launch changes what they can defend. The tender they were never going to run becomes a tender.
Objectives: grow at 0.4, stay solvent at 0.35, keep the board and the team with them at 0.25. Acts on the churn trigger, twice a quarter at most.
Objective: hit the number at weight 0.55. Three months under quota with discount authority turns into field discounting at half to all of the authority you granted, which is a price cut nobody decided to make. With no authority granted, this never fires.
Objective: work somewhere that still looks stable, weight 0.3. They do not react to a competitor directly. They react to the utilisation and morale that follow from losing work.
The ledger
| Assumption | What it does here | Why to check it first |
|---|---|---|
| Competitor strength | Multiplies the launch score for each rival and weights the market average | Read from a competitor file if you uploaded one. Otherwise it is derived from how important the graph thinks they are |
| Competitor price level | The denominator in every price comparison a customer makes | If you do not know what they charge, this is a default and the whole scenario is standing on it |
| Perceived quality | Your side of the ratio, and it does not move in this scenario | Usually estimated from reviews. Worth replacing with something you can defend |
| Cost to a customer of switching | Damps the price pressure and adds to a named account's reason to stay | The single biggest reason a good competitor product does not take your base |
| Monthly customer churn | The base that the deviation multiplies | Everything in this scenario is a multiple of this number |
| Customer satisfaction and referral strength | Feeds retention now and new business later | The slow half of the damage runs through here |
| Price elasticity of demand | Scales both the churn response and the new business response to their price cut | Widest band in the ledger |
Sensitivity measures which of these is actually moving the answer, rather than asserting it. Run it before you argue about any of them.
Honestly
This scenario is a good early warning system and a poor description of a product launch.
The largest gap is worth stating plainly: a better rival product raises your churn and loses you renewals, but it does not directly reduce your win rate on new business. New logos respond to your own absolute quality and to the price ratio, not to the quality ratio. If the real threat is that you stop winning deals you used to win, this scenario understates it.
The earliest month you could know
How much better their offer looks on the things your customers weigh, on a scale from nought to one, before it is scaled by how strong that competitor already is. It is a deliberately blunt number. Set it by asking what share of your own customers would say the new thing is better, and then put the band wide.
Run it both ways. The scenario as it ships is the do nothing case, which is the honest baseline for a decision about responding.
Then copy it, add your response as a second lever, and compare the two side by side against the same baseline on the same seeds. The difference is the value of responding, which is the number you actually wanted.
Because nothing happened to you. Quality in the model responds to strain, headcount, what you spend on the product and anything a scenario broke on purpose. A competitor getting better does not make you worse, it makes you relatively worse, and the model keeps those two things separate on purpose.
Competitors are created anyway, because a market with no rivals in it is not a market. They are marked inferred, drawn faded in the graph and listed as industry defaults in the ledger. The scenario still runs and the brief tells you how much of the answer is standing on invented rivals.
Next
The cheapest version of this scenario is the one you run in the quarter before it happens, when there is still time for the answer to change a retention budget.