Sales attribution: who really sold — the app, the dealer, the call centre or the campaign

Test

Take a specific sale. The customer signed up for a smartphone instalment plan. The app marks it as “sold via app”. The dealer records the contract as their sale. A campaign ran a week before purchase and the customer was in its target — marketing records it. The contact centre consulted the customer last week — they have a pre-sale interaction.

Ask all four who sold. Each will answer confidently — my channel. Add up the reports across channels and compare with the actual total — the sum exceeds the real number by a factor of 1.5 to 2.5. That is double and triple attribution.

If your organisation does not have a clear attribution model and each team writes into its report whatever passed through its system, channel budget allocation is sitting on a distorted base. This is not a theoretical concern — it is the reason channel-budgeting often delivers worse than expected.

Why this happens

Attribution in telecom historically grew along channels rather than along the customer journey. Each channel was instrumented by its own vendor with its own definition of conversion. Billing as the master never had channel attribution — it recorded only the final transaction.

Over time each channel learned to claim credit. If the customer clicked an email, landed on a page, then left and signed up at the office a week later, the email channel can claim that conversion. The dealer also claims — they signed the contract. Marketing claims — they ran the campaign. The contact centre claims — they answered the question.

Without rules everyone takes credit and the sum does not match reality. Financial reports show only the final transaction and do not see which channel actually drove it.

Where this creates commercial problems

Budget reallocation toward channels that work less than they appear. If digital “shows” high conversion in its own attribution, while in reality 60% of those conversions happened thanks to retail, budget shifts to digital and retail is cut. After the cut, total sales fall, and nobody understands why.

Misaligned team incentives. The dealer network beats KPI by claiming conversions that flowed through the app. Digital also beats KPI. Everyone is pleased, but total revenue is not growing.

Inability to evaluate new channels correctly. When a new initiative launches (say, a partner channel), its effectiveness is compared against existing channel cost-per-acquisition. If existing costs were measured on double-counted attribution, the comparison is distorted and the new channel may be wrongly judged “expensive”.

Strategic decisions on channel mix taken on air. The board discusses “increase digital investment” or “optimise retail” based on reports that do not reflect reality.

The attribution models that exist

Each model has its zone of applicability.

Last-touch. Credit goes to the last channel before conversion. If the customer was at the office before signing — credit to the dealer. Simple but ignores the long path the customer took to the decision.

First-touch. Credit goes to the first contact channel. If the customer first heard about the service via online ad — credit to marketing. Ignores the role of downstream channels in actual conversion.

Linear. Credit is split equally across all channels in the journey. Does not reflect real causality.

Time-decay. Channels closer to conversion get more credit than distant ones. Recognises recency, but still an approximation.

Position-based. First touch and last touch get 40% each, the rest is distributed across intermediate. A hybrid commonly used.

Data-driven (algorithmic). The model learns from historical data which channel combinations actually drove conversions and distributes credit accordingly. Harder to deliver, but reflective of reality.

Which model is right depends on organisation structure. The key is to pick one and apply it consistently.

What usually gets in the way

Teams defending their credit. When the organisation moves to more accurate attribution, some teams’ numbers fall — they stop claiming conversions that were not really theirs. Politically painful.

No connected customer journey view technically. Each channel keeps its own log, disconnected from others. Without unified journey, attribution is mathematically impossible.

Billing as the master sale source has no channel attribution. The final transaction is recorded, but without journey metadata.

Vendor-driven loyalty to channel-tools. Each channel uses its own analytics, and each tool argues that its channel works best. Vendors are not impartial.

Inability to separate organic from paid. Some sales happen because the customer decided themselves. Many models attribute these to the last contact channel, and the real paid effect gets overstated.

What real attribution requires

A unified customer journey view. All customer touchpoints — website visit, app open, contact centre call, retail visit, email click, ad view — connected through a stable customer ID. Without this, attribution does not work mathematically.

A single source of truth for conversion. Billing as the final sale registry. Every other system shows upstream activity, but conversion is attributable only to the final billing record.

Choice and lock-in of an attribution model. A C-suite decision, not a per-team one. Applied uniformly across channels.

Holdout testing. Part of the target audience does not receive a touch from the chosen channel, and a comparison of conversion rates shows the real incremental effect. A slow but honest measurement.

Decoupling team KPIs from raw conversion claims. Instead of “team claims its own conversions”, a shared pool is split by the agreed model.

When attribution reform is not a priority

If the organisation has just emerged from major restructuring and teams are still adjusting — adding attribution change adds conflict.

If customer journey tracking does not exist technically (no unified ID, no journey log), attribution reform hits the data foundation. Foundation first.

If competitive intensity does not allow time for measurement (everything moves fast, holdout testing cannot keep up), precise attribution is lower priority.

If the CFO does not have authority to demand attribution discipline, reform is blocked by political team disputes.

If existing channel attribution is hard-wired into commercial contracts with partners, rebuilding requires renegotiation, which takes time.

Discussion points for the committee

What attribution model is in use today? If the answer is “each channel its own” — that is the problem.

What is the total of sales claims across channels versus actual revenue? If the gap is more than 30%, attribution is definitely double-counting.

Which channel has the strongest numbers in its own reports versus skeptical numbers in the CFO’s view? That is a signal of overclaim.

What is the data foundation for unified customer journey? If absent — that is the first block.

Is C-level ready to demand attribution discipline even when some teams’ numbers fall? Without commitment the reform does not start.

How SamaraliSoft can help

Sales Attribution Reset — audit of current attribution per channel, identification of overlap and double-counting, choice of an attribution model fitting the operator, design of a unified customer journey view, organisational realignment of team KPIs, and a 90-day pilot of new attribution with measurement.

Related reading

• /en/insights/telecom-subscriber-intelligence-operating-model/ — operating model
• /en/insights/telecom-app-monetization/ — frequency and engagement
• /en/insights/telecom-arpu-bundles-devices/ — where ARPU grows
• /en/insights/telecom-offer-fatigue/ — why mass campaigns do not work

Sources

• Digital Government Portal — digital infrastructure UZ
• UzDaily — Central Asia telecom market
• Ministry of Digital Technologies — telecom in UZ
• MNP.UZ — number portability